# Which benchmarks predict which

Calibration · https://benchgap.net/calibration · data retrieved 2026-11-21

For each ordered pair of same-capability benchmarks with at least 5 shared models, several monotone curves are fitted and the one with the lowest leave-one-out error is kept, if it passes the quality gate (R² ≥ 0.3, error ≤ 15 pp).

| Calibration | Curve | Models | Error | Estimates |
|---|---|---|---|---|
| [BenchCAD Vision2Code (no tools) → BenchCAD Vision2Code (tools)](https://benchgap.net/calibration/benchcadvision2code/current/to/benchcadvision2codewithtools/current.md) | offset logistic | 5 | 0.0 pp | 0 |
| [Chartography (tools) → BenchCAD Vision2Code (tools)](https://benchgap.net/calibration/chartographywithtools/current/to/benchcadvision2codewithtools/current.md) | inverse Michaelis–Menten | 5 | 0.0 pp | 4 |
| [BenchCAD Vision2Code (tools) → BenchCAD Vision2Code (no tools)](https://benchgap.net/calibration/benchcadvision2codewithtools/current/to/benchcadvision2code/current.md) | linear | 5 | 0.1 pp | 1 |
| [Chartography (tools) → BenchCAD Vision2Code (no tools)](https://benchgap.net/calibration/chartographywithtools/current/to/benchcadvision2code/current.md) | Hill | 5 | 0.1 pp | 4 |
| [GMMLU → MILU](https://benchgap.net/calibration/gmmlu/current/to/milu/current.md) | Hill | 5 | 0.2 pp | 0 |
| [OfficeQA Pro → OfficeQA](https://benchgap.net/calibration/officeqapro/current/to/officeqa/current.md) | offset logistic | 5 | 0.2 pp | 6 |
| [AA-Omniscience Accuracy → STEMBench](https://benchgap.net/calibration/omniscienceaccuracy/current/to/stembench/current.md) | linear | 8 | 0.2 pp | 162 |
| [BioBasicsBench → GPQA](https://benchgap.net/calibration/biobasicsbench/current/to/gpqa/current.md) | Michaelis–Menten | 5 | 0.3 pp | 2 |
| [MILU → GMMLU](https://benchgap.net/calibration/milu/current/to/gmmlu/current.md) | linear | 5 | 0.3 pp | 0 |
| [OfficeQA → OfficeQA Pro](https://benchgap.net/calibration/officeqa/current/to/officeqapro/current.md) | linear | 5 | 0.3 pp | 0 |
| [HealthBench Hard → HealthBench (length-adjusted)](https://benchgap.net/calibration/healthbench-hard/current/to/healthbenchlengthadjusted/current.md) | Hill | 5 | 0.4 pp | 5 |
| [HealthBench (raw) → HealthBench Hard](https://benchgap.net/calibration/healthbench/current/to/healthbench-hard/current.md) | Hill | 5 | 0.4 pp | 0 |
| [MathVision → MathVision w/ Python](https://benchgap.net/calibration/mathvision/current/to/mathvisionpython/current.md) | Michaelis–Menten + offset | 5 | 0.5 pp | 12 |
| [MAXIFE → NOVA-63](https://benchgap.net/calibration/maxife/current/to/nova63/current.md) | linear | 5 | 0.5 pp | 0 |
| [MMMLU → MMLU-Redux](https://benchgap.net/calibration/mmmlu/current/to/mmluredux/current.md) | offset logistic | 7 | 0.5 pp | 7 |
| [MMLU-Pro (Arcee) → CritPt](https://benchgap.net/calibration/mmlu-pro-arcee/current/to/critpt/current.md) | offset logistic | 6 | 0.6 pp | 0 |
| [CritPt → BioMysteryBench (human-solvable)](https://benchgap.net/calibration/critpt/current/to/biomysterybenchhumansolvable/current.md) | Hill | 5 | 0.6 pp | 202 |
| [HealthBench (length-adjusted) → HealthBench Hard](https://benchgap.net/calibration/healthbenchlengthadjusted/current/to/healthbench-hard/current.md) | Hill | 5 | 0.6 pp | 2 |
| [MMLU-Pro → MMLU-Redux](https://benchgap.net/calibration/mmlu-pro/current/to/mmluredux/current.md) | Hill | 10 | 0.6 pp | 61 |
| [MCP Atlas → MCP-Tasks](https://benchgap.net/calibration/mcpatlas/current/to/mcptasks/current.md) | Hill | 5 | 0.6 pp | 14 |
| [RefCOCO (avg) → CC-OCR](https://benchgap.net/calibration/refcocoavg/current/to/ccocr/current.md) | linear | 5 | 0.7 pp | 5 |
| [GraphWalks BFS 256K–1M → ARC-AGI-1](https://benchgap.net/calibration/graphwalks-bfs256k1m/current/to/arcagi1/current.md) | offset logistic | 5 | 0.7 pp | 0 |
| [MathVision w/ Python → MathVision](https://benchgap.net/calibration/mathvisionpython/current/to/mathvision/current.md) | offset logistic | 5 | 0.7 pp | 1 |
| [HealthBench Hard → HealthBench (raw)](https://benchgap.net/calibration/healthbench-hard/current/to/healthbench/current.md) | Hill | 5 | 0.7 pp | 5 |
| [CritPt → MMMLU](https://benchgap.net/calibration/critpt/current/to/mmmlu/current.md) | Hill | 6 | 0.7 pp | 183 |
| [MMAnswerBench → AIME26](https://benchgap.net/calibration/mmanswerbench/current/to/aime2026/current.md) | linear | 10 | 0.7 pp | 0 |
| [MazeBench → DRACO](https://benchgap.net/calibration/mazebench/current/to/draco/current.md) | offset logistic | 5 | 0.7 pp | 50 |
| [FrontierCode 1.1 Extended → VulcanBench v3](https://benchgap.net/calibration/frontiercode11extended/current/to/vulcanbench%23vulcanbench-v3/current.md) | offset logistic | 5 | 0.7 pp | 3 |
| [AA-Omniscience Accuracy → AA MMLU-Pro](https://benchgap.net/calibration/omniscienceaccuracy/current/to/aa-mmlu-pro/current.md) | linear | 11 | 0.8 pp | 110 |
| [CritPt → HealthBench Professional (raw, Artificial Analysis)](https://benchgap.net/calibration/critpt/current/to/aa-health-bench-professional-raw/current.md) | Michaelis–Menten | 6 | 0.8 pp | 94 |
| [C-Eval → CritPt](https://benchgap.net/calibration/ceval/current/to/critpt/current.md) | linear | 6 | 0.8 pp | 4 |
| [SciCode → Vals LiveCodeBench](https://benchgap.net/calibration/scicode/current/to/valslivecodebench/current.md) | linear | 7 | 0.8 pp | 0 |
| [AA-Omniscience Accuracy → MMLU-Redux](https://benchgap.net/calibration/omniscienceaccuracy/current/to/mmluredux/current.md) | linear | 5 | 0.8 pp | 132 |
| [CharXiv → DynaMath](https://benchgap.net/calibration/charxiv/current/to/dynamath/current.md) | Hill | 6 | 0.8 pp | 43 |
| [LABBench2 → ARC-AGI-1](https://benchgap.net/calibration/labbench2/current/to/arcagi1/current.md) | Michaelis–Menten | 6 | 0.8 pp | 1 |
| [CWE-bench → Toolathlon-Verified](https://benchgap.net/calibration/cwe-bench-v2/current/to/toolathlonverified/current.md) | offset logistic | 5 | 0.8 pp | 15 |
| [MMMU-Pro w/ Python → MMMU-Pro](https://benchgap.net/calibration/mmmupropython/current/to/mmmu-pro/current.md) | Hill | 10 | 0.8 pp | 0 |
| [ARC-AGI-3 → HealthBench (length-adjusted)](https://benchgap.net/calibration/arcagi3/current/to/healthbenchlengthadjusted/current.md) | Hill | 5 | 0.9 pp | 17 |
| [MedXpertQA (Text) → GPQA Diamond](https://benchgap.net/calibration/medxpertqatext/current/to/gpqa-diamond/current.md) | offset logistic | 5 | 0.9 pp | 1 |
| [SuperGPQA → MMLU-Redux](https://benchgap.net/calibration/supergpqa/current/to/mmluredux/current.md) | Michaelis–Menten | 9 | 0.9 pp | 2 |
| [SWE Multilingual → MILU](https://benchgap.net/calibration/swe-bench-multilingual/current/to/milu/current.md) | Hill | 5 | 0.9 pp | 4 |
| [ARC-AGI-3 → GPQA](https://benchgap.net/calibration/arcagi3/current/to/gpqa/current.md) | linear | 12 | 0.9 pp | 10 |
| [BabyVision → MathVision](https://benchgap.net/calibration/babyvision/current/to/mathvision/current.md) | Michaelis–Menten | 9 | 0.9 pp | 3 |
| [HLE → C-Eval](https://benchgap.net/calibration/hle/current/to/ceval/current.md) | Michaelis–Menten | 6 | 0.9 pp | 29 |
| [CharXiv → AI2D_TEST](https://benchgap.net/calibration/charxiv/current/to/ai2dtest/current.md) | offset logistic | 7 | 0.9 pp | 42 |
| [AA-Omniscience Accuracy → MMLU](https://benchgap.net/calibration/omniscienceaccuracy/current/to/mmlu/current.md) | offset logistic | 10 | 0.9 pp | 146 |
| [CharXiv w/o tools → ParseBench-1](https://benchgap.net/calibration/charxivnotools/current/to/parse-bench-llama-index/current.md) | offset logistic | 6 | 1.0 pp | 12 |
| [Artificial Analysis Intelligence Index → C-Eval](https://benchgap.net/calibration/artificialanalysis/current/to/ceval/current.md) | linear | 5 | 1.0 pp | 132 |
| [PolyMath → MMLU-ProX](https://benchgap.net/calibration/polymath/current/to/mmluprox/current.md) | Hill | 5 | 1.0 pp | 1 |
| [MMMLU → C-Eval](https://benchgap.net/calibration/mmmlu/current/to/ceval/current.md) | Hill | 5 | 1.0 pp | 2 |
| [HMMT Feb 2025 → AIME26](https://benchgap.net/calibration/hmmtfeb2025/current/to/aime2026/current.md) | Michaelis–Menten | 7 | 1.0 pp | 4 |
| [SuperGPQA → C-Eval](https://benchgap.net/calibration/supergpqa/current/to/ceval/current.md) | Hill | 6 | 1.0 pp | 4 |
| [CharXiv → MathVision w/ Python](https://benchgap.net/calibration/charxiv/current/to/mathvisionpython/current.md) | Michaelis–Menten + offset | 6 | 1.0 pp | 9 |
| [MathVision → AI2D_TEST](https://benchgap.net/calibration/mathvision/current/to/ai2dtest/current.md) | linear | 5 | 1.0 pp | 2 |
| [OSWorld 2.0 → Toolathlon-Verified](https://benchgap.net/calibration/osworld2/current/to/toolathlonverified/current.md) | Michaelis–Menten | 5 | 1.0 pp | 16 |
| [Video-MME (w/o subtitle) → DynaMath](https://benchgap.net/calibration/videommenosub/current/to/dynamath/current.md) | offset logistic | 6 | 1.0 pp | 1 |
| [BenchCAD Vision2Code (no tools) → Chartography (tools)](https://benchgap.net/calibration/benchcadvision2code/current/to/chartographywithtools/current.md) | Michaelis–Menten | 5 | 1.1 pp | 0 |
| [MedXpertQA (MM) → MMMU-Pro](https://benchgap.net/calibration/medxpertqamm/current/to/mmmu-pro/current.md) | linear | 8 | 1.1 pp | 0 |
| [ExploitBench Internal Port → DAYJOB Healthcare](https://benchgap.net/calibration/exploitbenchinternalport/current/to/dayjob-healthcare/current.md) | offset logistic | 5 | 1.1 pp | 4 |
| [Artificial Analysis Intelligence Index → HealthBench Professional (raw, Artificial Analysis)](https://benchgap.net/calibration/artificialanalysis/current/to/aa-health-bench-professional-raw/current.md) | Michaelis–Menten | 6 | 1.1 pp | 66 |
| [BioBasicsBench → AA-GPQA Diamond](https://benchgap.net/calibration/biobasicsbench/current/to/aa-gpqa-diamond/current.md) | Michaelis–Menten | 5 | 1.1 pp | 2 |
| [CritPt → C-Eval](https://benchgap.net/calibration/critpt/current/to/ceval/current.md) | linear | 6 | 1.1 pp | 36 |
| [HLE-Verified → ARC-AGI-1](https://benchgap.net/calibration/hleverified/current/to/arcagi1/current.md) | offset logistic | 5 | 1.1 pp | 0 |
| [MathVision w/ Python → CharXiv](https://benchgap.net/calibration/mathvisionpython/current/to/charxiv/current.md) | Michaelis–Menten + offset | 6 | 1.1 pp | 0 |
| [Vals MMLU-Pro → AA MMLU-Pro](https://benchgap.net/calibration/vals-mmlu-pro/current/to/aa-mmlu-pro/current.md) | Michaelis–Menten + offset | 10 | 1.1 pp | 13 |
| [C-Eval → MMLU-Redux](https://benchgap.net/calibration/ceval/current/to/mmluredux/current.md) | Hill | 10 | 1.1 pp | 0 |
| [GPQA Diamond → MMMLU](https://benchgap.net/calibration/gpqa-diamond/current/to/mmmlu/current.md) | Hill | 5 | 1.1 pp | 30 |
| [ExploitBench → DAYJOB](https://benchgap.net/calibration/exploitbench/current/to/dayjob/current.md) | inverse Michaelis–Menten | 5 | 1.1 pp | 13 |
| [C-Eval → MMMLU](https://benchgap.net/calibration/ceval/current/to/mmmlu/current.md) | linear | 5 | 1.2 pp | 0 |
| [LABBench2 → HLE-Verified](https://benchgap.net/calibration/labbench2/current/to/hleverified/current.md) | offset logistic | 6 | 1.2 pp | 1 |
| [Artificial Analysis Intelligence Index → AA MMLU-Pro](https://benchgap.net/calibration/artificialanalysis/current/to/aa-mmlu-pro/current.md) | Michaelis–Menten + offset | 12 | 1.2 pp | 16 |
| [GPQA → AA MMLU-Pro](https://benchgap.net/calibration/gpqa/current/to/aa-mmlu-pro/current.md) | linear | 5 | 1.2 pp | 46 |
| [AA Agentic Index → AA Harvey LAB v1.0](https://benchgap.net/calibration/aa-agentic-index/current/to/aa-harvey-lab/current.md) | Hill | 7 | 1.2 pp | 62 |
| [CursorBench 4.0 → AA Coding Index](https://benchgap.net/calibration/cursor-bench40/current/to/aa-coding-index/current.md) | linear | 9 | 1.2 pp | 12 |
| [GPQA → C-Eval](https://benchgap.net/calibration/gpqa/current/to/ceval/current.md) | Michaelis–Menten | 11 | 1.2 pp | 46 |
| [BioMysteryBench (human-difficult) → BioMysteryBench (human-solvable)](https://benchgap.net/calibration/biomysterybenchhumandifficult/current/to/biomysterybenchhumansolvable/current.md) | offset logistic | 5 | 1.2 pp | 0 |
| [Vals MMLU-Pro → MMLU](https://benchgap.net/calibration/vals-mmlu-pro/current/to/mmlu/current.md) | Hill | 8 | 1.2 pp | 22 |
| [ProgramBench → VulcanBench Frontier v4](https://benchgap.net/calibration/programbench/current/to/vulcanbench/current.md) | offset logistic | 6 | 1.2 pp | 15 |
| [JobBench → WideResearch](https://benchgap.net/calibration/jobbench/current/to/wideresearch/current.md) | linear | 5 | 1.2 pp | 32 |
| [MMMU-Pro → CharXiv w/o tools](https://benchgap.net/calibration/mmmu-pro/current/to/charxivnotools/current.md) | offset logistic | 9 | 1.2 pp | 50 |
| [HealthBench Professional → Vals GPQA Diamond](https://benchgap.net/calibration/healthbenchprofessional/current/to/vals-gpqa-diamond/current.md) | offset logistic | 6 | 1.2 pp | 15 |
| [HMMT Feb 2025 → HMMT Feb 2026](https://benchgap.net/calibration/hmmtfeb2025/current/to/hmmtfeb2026/current.md) | Michaelis–Menten + offset | 7 | 1.2 pp | 4 |
| [MMMU → Video-MME (w/o subtitle)](https://benchgap.net/calibration/mmmu/current/to/videommenosub/current.md) | Hill | 9 | 1.2 pp | 23 |
| [JobBench → AA EnterpriseOps-Gym](https://benchgap.net/calibration/jobbench/current/to/aa-enterprise-ops-gym/current.md) | Michaelis–Menten | 5 | 1.2 pp | 32 |
| [AA-GPQA Diamond → C-Eval](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/ceval/current.md) | Hill | 7 | 1.2 pp | 9 |
| [DRACO → Toolathlon-Verified](https://benchgap.net/calibration/draco/current/to/toolathlonverified/current.md) | offset logistic | 7 | 1.2 pp | 2 |
| [GPQA Diamond → GPQA](https://benchgap.net/calibration/gpqa-diamond/current/to/gpqa/current.md) | linear | 56 | 1.2 pp | 17 |
| [GPQA → GPQA Diamond](https://benchgap.net/calibration/gpqa/current/to/gpqa-diamond/current.md) | linear | 56 | 1.2 pp | 76 |
| [HealthBench Professional (raw) → HealthBench Hard](https://benchgap.net/calibration/healthbenchprofessionalraw/current/to/healthbench-hard/current.md) | inverse Michaelis–Menten | 5 | 1.2 pp | 2 |
| [HealthBench (raw) → HealthBench Professional (raw)](https://benchgap.net/calibration/healthbench/current/to/healthbenchprofessionalraw/current.md) | Michaelis–Menten | 8 | 1.2 pp | 0 |
| [Artificial Analysis Intelligence Index → MMLU-Redux](https://benchgap.net/calibration/artificialanalysis/current/to/mmluredux/current.md) | Michaelis–Menten | 6 | 1.2 pp | 17 |
| [PostTrainBench v1 → PostTrainBench](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1/current/to/post-train-bench12/current.md) | linear | 6 | 1.2 pp | 7 |
| [MMLU-Redux → MMLU-Pro](https://benchgap.net/calibration/mmluredux/current/to/mmlu-pro/current.md) | inverse Michaelis–Menten | 10 | 1.2 pp | 0 |
| [LABBench2 → AA-GPQA Diamond](https://benchgap.net/calibration/labbench2/current/to/aa-gpqa-diamond/current.md) | Michaelis–Menten | 6 | 1.2 pp | 1 |
| [FrontierMath v2 (Tier 4) → AIME26](https://benchgap.net/calibration/frontiermathv2tier4/current/to/aime2026/current.md) | offset logistic | 5 | 1.2 pp | 30 |
| [HealthBench Professional → GPQA](https://benchgap.net/calibration/healthbenchprofessional/current/to/gpqa/current.md) | offset logistic | 7 | 1.2 pp | 9 |
| [HealthBench Professional → GPQA Diamond](https://benchgap.net/calibration/healthbenchprofessional/current/to/gpqa-diamond/current.md) | offset logistic | 7 | 1.2 pp | 13 |
| [MMLU-ProX → NOVA-63](https://benchgap.net/calibration/mmluprox/current/to/nova63/current.md) | Michaelis–Menten | 7 | 1.3 pp | 5 |
| [MathVision → CountBench](https://benchgap.net/calibration/mathvision/current/to/countbench/current.md) | offset logistic | 6 | 1.3 pp | 19 |
| [Mystery Game Puzzles → Cookie-Bench](https://benchgap.net/calibration/mysterygamepuzzles/current/to/cookie-bench/current.md) | Hill | 5 | 1.3 pp | 71 |
| [AIME26 → MMAnswerBench](https://benchgap.net/calibration/aime2026/current/to/mmanswerbench/current.md) | inverse Michaelis–Menten | 10 | 1.3 pp | 18 |
| [PostTrainBench → PostTrainBench v1](https://benchgap.net/calibration/post-train-bench12/current/to/posttrainbench%23posttrainbench-v1/current.md) | linear | 6 | 1.3 pp | 17 |
| [CharXiv w/o tools → MathVision w/ Python](https://benchgap.net/calibration/charxivnotools/current/to/mathvisionpython/current.md) | Michaelis–Menten | 5 | 1.3 pp | 1 |
| [PolyMath → NOVA-63](https://benchgap.net/calibration/polymath/current/to/nova63/current.md) | Michaelis–Menten | 5 | 1.3 pp | 1 |
| [SWE Multilingual → GMMLU](https://benchgap.net/calibration/swe-bench-multilingual/current/to/gmmlu/current.md) | Hill | 5 | 1.3 pp | 4 |
| [Video-MME (with subtitle) → AI2D_TEST](https://benchgap.net/calibration/videommewithsub/current/to/ai2dtest/current.md) | Michaelis–Menten + offset | 6 | 1.3 pp | 2 |
| [OSWorld 2.0 → AA Harvey LAB v1.0](https://benchgap.net/calibration/osworld2/current/to/aa-harvey-lab/current.md) | linear | 5 | 1.3 pp | 7 |
| [PostTrainBench v1 → PostTrainBench v1.1](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1/current/to/posttrainbench%23posttrainbench-v1-1/current.md) | Hill | 6 | 1.3 pp | 7 |
| [HMMT Feb 2026 → AIME26](https://benchgap.net/calibration/hmmtfeb2026/current/to/aime2026/current.md) | Hill | 22 | 1.3 pp | 3 |
| [AIME 2024 → MATH-500](https://benchgap.net/calibration/aime2024/current/to/math-500/current.md) | Michaelis–Menten | 15 | 1.3 pp | 5 |
| [MMLU-Pro → C-Eval](https://benchgap.net/calibration/mmlu-pro/current/to/ceval/current.md) | Hill | 7 | 1.3 pp | 2 |
| [MMMU-Pro → Video-MME (w/o subtitle)](https://benchgap.net/calibration/mmmu-pro/current/to/videommenosub/current.md) | Michaelis–Menten | 8 | 1.3 pp | 40 |
| [SkillsBench → WideResearch](https://benchgap.net/calibration/skills-bench/current/to/wideresearch/current.md) | Hill | 6 | 1.3 pp | 1 |
| [DeepPlanning → MCP-Tasks](https://benchgap.net/calibration/deepplanning/current/to/mcptasks/current.md) | Hill | 5 | 1.3 pp | 1 |
| [ARC-AGI-2 → HLE-Verified](https://benchgap.net/calibration/arc-agi-2/current/to/hleverified/current.md) | Hill | 5 | 1.3 pp | 17 |
| [SWE Multilingual → MMLU-ProX](https://benchgap.net/calibration/swe-bench-multilingual/current/to/mmluprox/current.md) | Hill | 8 | 1.3 pp | 37 |
| [AA-HLE → AA MMLU-Pro](https://benchgap.net/calibration/aa-hle/current/to/aa-mmlu-pro/current.md) | offset logistic | 13 | 1.3 pp | 32 |
| [Claw-Eval → ResearchClawBench](https://benchgap.net/calibration/claw-eval/current/to/researchclawbench/current.md) | Michaelis–Menten | 12 | 1.3 pp | 0 |
| [AA-HLE → BioMysteryBench (human-solvable)](https://benchgap.net/calibration/aa-hle/current/to/biomysterybenchhumansolvable/current.md) | Michaelis–Menten | 5 | 1.4 pp | 1 |
| [RealWorldQA → AI2D_TEST](https://benchgap.net/calibration/realworldqa/current/to/ai2dtest/current.md) | offset logistic | 7 | 1.4 pp | 5 |
| [FrontierCode 1.1 Extended → Vals SWE-bench](https://benchgap.net/calibration/frontiercode11extended/current/to/valsswebench/current.md) | Michaelis–Menten | 5 | 1.4 pp | 3 |
| [Decision Index Full Score → Artificial Analysis Decision Index](https://benchgap.net/calibration/decisionindex/current/to/aa-decision-index-v2/current.md) | offset logistic | 11 | 1.4 pp | 0 |
| [NOVA-63 → MMLU-ProX](https://benchgap.net/calibration/nova63/current/to/mmluprox/current.md) | Michaelis–Menten + offset | 7 | 1.4 pp | 0 |
| [ScreenSpot Pro → RealWorldQA](https://benchgap.net/calibration/screenspot-pro/current/to/realworldqa/current.md) | linear | 7 | 1.4 pp | 0 |
| [DynaMath → RealWorldQA](https://benchgap.net/calibration/dynamath/current/to/realworldqa/current.md) | Michaelis–Menten + offset | 7 | 1.4 pp | 0 |
| [ARC-AGI-3 → Vals GPQA Diamond](https://benchgap.net/calibration/arcagi3/current/to/vals-gpqa-diamond/current.md) | Michaelis–Menten | 13 | 1.4 pp | 4 |
| [CritPt → AA MMLU-Pro](https://benchgap.net/calibration/critpt/current/to/aa-mmlu-pro/current.md) | Michaelis–Menten | 12 | 1.4 pp | 3 |
| [Chess Puzzles → ExploitBench Internal Port](https://benchgap.net/calibration/chesspuzzles/current/to/exploitbenchinternalport/current.md) | offset logistic | 6 | 1.4 pp | 76 |
| [DeepPlanning → τ³-bench results](https://benchgap.net/calibration/deepplanning/current/to/tau3-bench/current.md) | Michaelis–Menten | 6 | 1.4 pp | 1 |
| [MMLU-Redux → MMMLU](https://benchgap.net/calibration/mmluredux/current/to/mmmlu/current.md) | offset logistic | 7 | 1.4 pp | 4 |
| [MCP-Tasks → τ³-bench results](https://benchgap.net/calibration/mcptasks/current/to/tau3-bench/current.md) | Michaelis–Menten | 5 | 1.4 pp | 0 |
| [ARC-AGI-1 → GPQA Diamond](https://benchgap.net/calibration/arcagi1/current/to/gpqa-diamond/current.md) | offset logistic | 15 | 1.4 pp | 16 |
| [MMLU-ProX → INCLUDE](https://benchgap.net/calibration/mmluprox/current/to/include/current.md) | Michaelis–Menten | 5 | 1.4 pp | 7 |
| [MMLU-Redux → C-Eval](https://benchgap.net/calibration/mmluredux/current/to/ceval/current.md) | linear | 10 | 1.4 pp | 1 |
| [CountBench → AI2D_TEST](https://benchgap.net/calibration/countbench/current/to/ai2dtest/current.md) | linear | 6 | 1.4 pp | 0 |
| [FrontierMath v2 (Tiers 1-3) → AIME26](https://benchgap.net/calibration/frontiermathv2tiers13/current/to/aime2026/current.md) | Michaelis–Menten | 9 | 1.4 pp | 31 |
| [LABBench2 → Vals MMLU-Pro](https://benchgap.net/calibration/labbench2/current/to/vals-mmlu-pro/current.md) | Michaelis–Menten + offset | 6 | 1.4 pp | 1 |
| [ExploitBench → DAYJOB Healthcare](https://benchgap.net/calibration/exploitbench/current/to/dayjob-healthcare/current.md) | offset logistic | 5 | 1.5 pp | 9 |
| [ARC-AGI-2 → GPQA Diamond](https://benchgap.net/calibration/arc-agi-2/current/to/gpqa-diamond/current.md) | linear | 19 | 1.5 pp | 2 |
| [CC-OCR → OmniDocBench 1.5](https://benchgap.net/calibration/ccocr/current/to/omnidocbench15/current.md) | Michaelis–Menten | 6 | 1.5 pp | 2 |
| [FrontierMath v2 (Tiers 1-3) → HMMT Nov 2025](https://benchgap.net/calibration/frontiermathv2tiers13/current/to/hmmtnov2025/current.md) | Michaelis–Menten | 6 | 1.5 pp | 65 |
| [FrontierMath v2 (Tier 4) → HMMT Feb 2026](https://benchgap.net/calibration/frontiermathv2tier4/current/to/hmmtfeb2026/current.md) | linear | 6 | 1.5 pp | 33 |
| [DAYJOB Healthcare → DAYJOB](https://benchgap.net/calibration/dayjob-healthcare/current/to/dayjob/current.md) | Michaelis–Menten | 19 | 1.5 pp | 0 |
| [MLVU (M-Avg) → OmniDocBench 1.5](https://benchgap.net/calibration/mlvuavg/current/to/omnidocbench15/current.md) | offset logistic | 8 | 1.5 pp | 0 |
| [MCP-Tasks → MCP Atlas](https://benchgap.net/calibration/mcptasks/current/to/mcpatlas/current.md) | inverse Michaelis–Menten | 5 | 1.5 pp | 0 |
| [AA-HLE → C-Eval](https://benchgap.net/calibration/aa-hle/current/to/ceval/current.md) | Michaelis–Menten | 7 | 1.5 pp | 5 |
| [MathVision → MLVU (M-Avg)](https://benchgap.net/calibration/mathvision/current/to/mlvuavg/current.md) | linear | 8 | 1.5 pp | 17 |
| [RealWorldQA → DynaMath](https://benchgap.net/calibration/realworldqa/current/to/dynamath/current.md) | offset logistic | 7 | 1.5 pp | 2 |
| [CharXiv → RealWorldQA](https://benchgap.net/calibration/charxiv/current/to/realworldqa/current.md) | Michaelis–Menten | 12 | 1.5 pp | 29 |
| [AA-MMMU-Pro → MMMU-Pro](https://benchgap.net/calibration/aa-mmmu-pro/current/to/mmmu-pro/current.md) | offset logistic | 39 | 1.5 pp | 70 |
| [HMMT Feb 2025 → HMMT Nov 2025](https://benchgap.net/calibration/hmmtfeb2025/current/to/hmmtnov2025/current.md) | linear | 9 | 1.5 pp | 6 |
| [MMMU-Pro → AA-MMMU-Pro](https://benchgap.net/calibration/mmmu-pro/current/to/aa-mmmu-pro/current.md) | linear | 39 | 1.5 pp | 19 |
| [ExploitGym → Hermes Bench](https://benchgap.net/calibration/exploitgym/current/to/hermesbench/current.md) | Michaelis–Menten | 6 | 1.5 pp | 17 |
| [Agents' Last Exam → Hermes Bench](https://benchgap.net/calibration/agentslastexam/current/to/hermesbench/current.md) | Hill | 7 | 1.5 pp | 9 |
| [CharXiv → MLVU (M-Avg)](https://benchgap.net/calibration/charxiv/current/to/mlvuavg/current.md) | linear | 9 | 1.5 pp | 26 |
| [CharXiv → VideoMMMU](https://benchgap.net/calibration/charxiv/current/to/videommmu/current.md) | Hill | 10 | 1.5 pp | 40 |
| [GPQA Diamond → Vals MMLU-Pro](https://benchgap.net/calibration/gpqa-diamond/current/to/vals-mmlu-pro/current.md) | Michaelis–Menten + offset | 25 | 1.6 pp | 36 |
| [HLE-Verified → AA-GPQA Diamond](https://benchgap.net/calibration/hleverified/current/to/aa-gpqa-diamond/current.md) | offset logistic | 6 | 1.6 pp | 0 |
| [CC-OCR → CountBench](https://benchgap.net/calibration/ccocr/current/to/countbench/current.md) | linear | 5 | 1.6 pp | 2 |
| [ARC-AGI-3 → Vals MMLU-Pro](https://benchgap.net/calibration/arcagi3/current/to/vals-mmlu-pro/current.md) | Michaelis–Menten | 13 | 1.6 pp | 5 |
| [Vals GPQA Diamond → AA MMLU-Pro](https://benchgap.net/calibration/vals-gpqa-diamond/current/to/aa-mmlu-pro/current.md) | linear | 10 | 1.6 pp | 0 |
| [ERQA → DynaMath](https://benchgap.net/calibration/erqa/current/to/dynamath/current.md) | Hill | 6 | 1.6 pp | 1 |
| [AA-GPQA Diamond → AA MMLU-Pro](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/aa-mmlu-pro/current.md) | Hill | 13 | 1.6 pp | 0 |
| [BrowseComp → LifeVal](https://benchgap.net/calibration/browsecomp/current/to/lifeval/current.md) | Hill | 5 | 1.6 pp | 56 |
| [MathVision → CharXiv w/o tools](https://benchgap.net/calibration/mathvision/current/to/charxivnotools/current.md) | Hill | 9 | 1.6 pp | 2 |
| [ERQA → RealWorldQA](https://benchgap.net/calibration/erqa/current/to/realworldqa/current.md) | Michaelis–Menten | 11 | 1.6 pp | 0 |
| [V* → CountBench](https://benchgap.net/calibration/vstar/current/to/countbench/current.md) | linear | 6 | 1.6 pp | 1 |
| [DynaMath → CharXiv](https://benchgap.net/calibration/dynamath/current/to/charxiv/current.md) | Hill | 6 | 1.6 pp | 1 |
| [CyberGym → WideResearch](https://benchgap.net/calibration/cybergym/current/to/wideresearch/current.md) | Hill | 5 | 1.6 pp | 15 |
| [HealthBench Professional (raw) → HealthBench (raw)](https://benchgap.net/calibration/healthbenchprofessionalraw/current/to/healthbench/current.md) | inverse Michaelis–Menten | 8 | 1.6 pp | 2 |
| [CursorBench 3.2 → FrontierCode 1.1 Main](https://benchgap.net/calibration/cursor-bench32/current/to/frontiercode/current.md) | inverse Michaelis–Menten | 5 | 1.6 pp | 13 |
| [CWE-bench → Hermes Bench](https://benchgap.net/calibration/cwe-bench-v2/current/to/hermesbench/current.md) | linear | 7 | 1.6 pp | 7 |
| [HMMT Nov 2025 → AIME26](https://benchgap.net/calibration/hmmtnov2025/current/to/aime2026/current.md) | Michaelis–Menten | 10 | 1.6 pp | 1 |
| [FrontierCode 1.1 Main → CursorBench 3.2](https://benchgap.net/calibration/frontiercode/current/to/cursor-bench32/current.md) | Michaelis–Menten | 5 | 1.6 pp | 9 |
| [CursorBench 3.2 → FrontierCode 1.1 Extended](https://benchgap.net/calibration/cursor-bench32/current/to/frontiercode11extended/current.md) | offset logistic | 5 | 1.6 pp | 11 |
| [ARC-AGI-2 → GPQA](https://benchgap.net/calibration/arc-agi-2/current/to/gpqa/current.md) | linear | 21 | 1.6 pp | 11 |
| [BenchCAD Vision2Code (tools) → Chartography (tools)](https://benchgap.net/calibration/benchcadvision2codewithtools/current/to/chartographywithtools/current.md) | Hill | 5 | 1.6 pp | 1 |
| [We-Math → AI2D_TEST](https://benchgap.net/calibration/wemath/current/to/ai2dtest/current.md) | Michaelis–Menten | 5 | 1.6 pp | 0 |
| [ApprenticeBench → Toolathlon](https://benchgap.net/calibration/apprenticebench/current/to/toolathlon/current.md) | linear | 5 | 1.7 pp | 11 |
| [HLE w/ tools → AA Harvey LAB v1.0](https://benchgap.net/calibration/hlewithtools/current/to/aa-harvey-lab/current.md) | Michaelis–Menten | 5 | 1.7 pp | 22 |
| [OSWorld 2.1 Offline → GDPval-AA](https://benchgap.net/calibration/osworld-2-1-offline/current/to/gdpvalaanormalized/current.md) | Hill | 6 | 1.7 pp | 5 |
| [CursorBench 4.0 → PostTrainBench v1.1](https://benchgap.net/calibration/cursor-bench40/current/to/posttrainbench%23posttrainbench-v1-1/current.md) | offset logistic | 5 | 1.7 pp | 16 |
| [HealthBench Hard → HealthBench Professional (raw)](https://benchgap.net/calibration/healthbench-hard/current/to/healthbenchprofessionalraw/current.md) | Michaelis–Menten | 5 | 1.7 pp | 5 |
| [HealthBench (length-adjusted) → HealthBench Professional](https://benchgap.net/calibration/healthbenchlengthadjusted/current/to/healthbenchprofessional/current.md) | Hill | 8 | 1.7 pp | 0 |
| [DynaMath → Video-MME (with subtitle)](https://benchgap.net/calibration/dynamath/current/to/videommewithsub/current.md) | offset logistic | 7 | 1.7 pp | 0 |
| [DynaMath → AI2D_TEST](https://benchgap.net/calibration/dynamath/current/to/ai2dtest/current.md) | Michaelis–Menten | 5 | 1.7 pp | 0 |
| [AA-MMMU-Pro → Video-MME (w/o subtitle)](https://benchgap.net/calibration/aa-mmmu-pro/current/to/videommenosub/current.md) | Hill | 6 | 1.7 pp | 65 |
| [ARC-AGI-2 → ARC-AGI-1](https://benchgap.net/calibration/arc-agi-2/current/to/arcagi1/current.md) | Michaelis–Menten + offset | 41 | 1.7 pp | 6 |
| [ARC-AGI-1 → LABBench2](https://benchgap.net/calibration/arcagi1/current/to/labbench2/current.md) | Hill | 6 | 1.7 pp | 18 |
| [HealthBench Hard → GPQA](https://benchgap.net/calibration/healthbench-hard/current/to/gpqa/current.md) | linear | 6 | 1.7 pp | 0 |
| [GPQA → MMMLU](https://benchgap.net/calibration/gpqa/current/to/mmmlu/current.md) | Michaelis–Menten | 14 | 1.7 pp | 31 |
| [PhysicianBench → HealthBench Professional (raw, Artificial Analysis)](https://benchgap.net/calibration/physicianbench/current/to/aa-health-bench-professional-raw/current.md) | linear | 5 | 1.7 pp | 0 |
| [CursorBench 3.1 → AA-SciCode](https://benchgap.net/calibration/cursor-bench31/current/to/aa-sci-code/current.md) | linear | 5 | 1.7 pp | 2 |
| [AA-GPQA Diamond → MMMLU](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/mmmlu/current.md) | Michaelis–Menten | 9 | 1.7 pp | 13 |
| [MMMU-Pro → MMMU](https://benchgap.net/calibration/mmmu-pro/current/to/mmmu/current.md) | linear | 19 | 1.7 pp | 13 |
| [GPQA → MMLU-Redux](https://benchgap.net/calibration/gpqa/current/to/mmluredux/current.md) | Michaelis–Menten + offset | 16 | 1.7 pp | 24 |
| [NYT Connections → FailureSensorIQ](https://benchgap.net/calibration/nyt-connections/current/to/failuresensoriq/current.md) | linear | 8 | 1.7 pp | 80 |
| [HLE w/ tools → Hermes Bench](https://benchgap.net/calibration/hlewithtools/current/to/hermesbench/current.md) | linear | 8 | 1.8 pp | 18 |
| [RealWorldQA → OmniDocBench 1.5](https://benchgap.net/calibration/realworldqa/current/to/omnidocbench15/current.md) | Hill | 9 | 1.8 pp | 5 |
| [HLE-Verified → LABBench2](https://benchgap.net/calibration/hleverified/current/to/labbench2/current.md) | offset logistic | 6 | 1.8 pp | 0 |
| [ARC-AGI-1 → GPQA](https://benchgap.net/calibration/arcagi1/current/to/gpqa/current.md) | linear | 19 | 1.8 pp | 1 |
| [Vals GPQA Diamond → GPQA Diamond](https://benchgap.net/calibration/vals-gpqa-diamond/current/to/gpqa-diamond/current.md) | Michaelis–Menten | 25 | 1.8 pp | 19 |
| [GDP.pdf → Hermes Bench](https://benchgap.net/calibration/aa-gdp-pdf/current/to/hermesbench/current.md) | offset logistic | 6 | 1.8 pp | 4 |
| [VideoMMMU → Video-MME (w/o subtitle)](https://benchgap.net/calibration/videommmu/current/to/videommenosub/current.md) | Michaelis–Menten + offset | 8 | 1.8 pp | 0 |
| [Vals MMLU-Pro → LABBench2](https://benchgap.net/calibration/vals-mmlu-pro/current/to/labbench2/current.md) | Michaelis–Menten + offset | 6 | 1.8 pp | 50 |
| [DAYJOB → DAYJOB Healthcare](https://benchgap.net/calibration/dayjob/current/to/dayjob-healthcare/current.md) | inverse Michaelis–Menten | 19 | 1.8 pp | 0 |
| [AA MMLU-Pro → Vals MMLU-Pro](https://benchgap.net/calibration/aa-mmlu-pro/current/to/vals-mmlu-pro/current.md) | inverse Michaelis–Menten | 10 | 1.8 pp | 3 |
| [INCLUDE → MMLU-ProX](https://benchgap.net/calibration/include/current/to/mmluprox/current.md) | Michaelis–Menten + offset | 5 | 1.8 pp | 0 |
| [MKA → Toolathlon-Verified](https://benchgap.net/calibration/mka/current/to/toolathlonverified/current.md) | Hill | 5 | 1.8 pp | 3 |
| [ParseBench-1 → CharXiv w/o tools](https://benchgap.net/calibration/parse-bench-llama-index/current/to/charxivnotools/current.md) | Michaelis–Menten | 6 | 1.8 pp | 16 |
| [C-Eval → MMLU-Pro](https://benchgap.net/calibration/ceval/current/to/mmlu-pro/current.md) | Hill | 7 | 1.8 pp | 0 |
| [AIME26 → AIME 2025](https://benchgap.net/calibration/aime2026/current/to/aime2025/current.md) | Hill | 7 | 1.8 pp | 27 |
| [CharXiv w/o tools → RealWorldQA](https://benchgap.net/calibration/charxivnotools/current/to/realworldqa/current.md) | Hill | 5 | 1.8 pp | 2 |
| [AA-GPQA Diamond → LABBench2](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/labbench2/current.md) | inverse Michaelis–Menten | 6 | 1.8 pp | 16 |
| [HMMT Feb 2026 → HMMT Feb 2025](https://benchgap.net/calibration/hmmtfeb2026/current/to/hmmtfeb2025/current.md) | Michaelis–Menten + offset | 7 | 1.8 pp | 6 |
| [CC-OCR → RefCOCO (avg)](https://benchgap.net/calibration/ccocr/current/to/refcocoavg/current.md) | linear | 5 | 1.8 pp | 2 |
| [CursorBench 4.0 → VulcanBench Frontier v4](https://benchgap.net/calibration/cursor-bench40/current/to/vulcanbench/current.md) | Hill | 10 | 1.8 pp | 8 |
| [HLE w/o tools → Vals MMLU-Pro](https://benchgap.net/calibration/hlenotools/current/to/vals-mmlu-pro/current.md) | Michaelis–Menten | 21 | 1.8 pp | 12 |
| [MCP Atlas → ResearchClawBench](https://benchgap.net/calibration/mcpatlas/current/to/researchclawbench/current.md) | linear | 13 | 1.8 pp | 29 |
| [PostTrainBench v1.1 → PostTrainBench v1](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1-1/current/to/posttrainbench%23posttrainbench-v1/current.md) | Hill | 6 | 1.8 pp | 2 |
| [CountBench → OmniDocBench 1.5](https://benchgap.net/calibration/countbench/current/to/omnidocbench15/current.md) | Michaelis–Menten + offset | 5 | 1.8 pp | 1 |
| [BinaryAudit → FailureSensorIQ](https://benchgap.net/calibration/binaryaudit/current/to/failuresensoriq/current.md) | Michaelis–Menten | 5 | 1.8 pp | 6 |
| [We-Math → CountBench](https://benchgap.net/calibration/wemath/current/to/countbench/current.md) | offset logistic | 6 | 1.8 pp | 0 |
| [ARC-AGI-3 → GPQA Diamond](https://benchgap.net/calibration/arcagi3/current/to/gpqa-diamond/current.md) | linear | 14 | 1.8 pp | 1 |
| [AA Agentic Index → ResearchClawBench](https://benchgap.net/calibration/aa-agentic-index/current/to/researchclawbench/current.md) | Michaelis–Menten | 11 | 1.9 pp | 41 |
| [FrontierCode 1.1 Main → Vals LiveCodeBench](https://benchgap.net/calibration/frontiercode/current/to/valslivecodebench/current.md) | linear | 9 | 1.9 pp | 8 |
| [ARC-AGI-2 → Vals MMLU-Pro](https://benchgap.net/calibration/arc-agi-2/current/to/vals-mmlu-pro/current.md) | Michaelis–Menten | 31 | 1.9 pp | 3 |
| [PolyMath → INCLUDE](https://benchgap.net/calibration/polymath/current/to/include/current.md) | offset logistic | 5 | 1.9 pp | 1 |
| [Artificial Analysis Decision Index v1.0 → HLE-MultiChoice (Decision Index)](https://benchgap.net/calibration/aa-decision-index/current/to/aa-decision-hle-multi-choice/current.md) | linear | 19 | 1.9 pp | 0 |
| [Decision Index Knowledge → Artificial Analysis Decision Index](https://benchgap.net/calibration/jevdecisionindex03knowledge/current/to/aa-decision-index-v2/current.md) | Michaelis–Menten | 11 | 1.9 pp | 10 |
| [We-Math → DynaMath](https://benchgap.net/calibration/wemath/current/to/dynamath/current.md) | Michaelis–Menten | 6 | 1.9 pp | 0 |
| [Vals GPQA Diamond → LABBench2](https://benchgap.net/calibration/vals-gpqa-diamond/current/to/labbench2/current.md) | Michaelis–Menten + offset | 6 | 1.9 pp | 1 |
| [VideoMMMU → MMMU](https://benchgap.net/calibration/videommmu/current/to/mmmu/current.md) | linear | 12 | 1.9 pp | 0 |
| [OSWorld 2.0 → BrowseComp](https://benchgap.net/calibration/osworld2/current/to/browsecomp/current.md) | offset logistic | 11 | 1.9 pp | 13 |
| [GDPval-AA → OSWorld 2.1 Offline](https://benchgap.net/calibration/gdpvalaanormalized/current/to/osworld-2-1-offline/current.md) | Hill | 6 | 1.9 pp | 67 |
| [Decision Index Language → WildGuardTest (Decision Index)](https://benchgap.net/calibration/jevdecisionindex03language/current/to/aa-decision-wild-guard-test/current.md) | linear | 11 | 1.9 pp | 0 |
| [HMMT Nov 2025 → HMMT Feb 2025](https://benchgap.net/calibration/hmmtnov2025/current/to/hmmtfeb2025/current.md) | linear | 9 | 1.9 pp | 2 |
| [Vision2Web → MathVision w/ Python](https://benchgap.net/calibration/vision2web/current/to/mathvisionpython/current.md) | Hill | 5 | 1.9 pp | 0 |
| [DAYJOB → ExploitBench](https://benchgap.net/calibration/dayjob/current/to/exploitbench/current.md) | Hill | 5 | 1.9 pp | 14 |
| [MathVision → Video-MME (with subtitle)](https://benchgap.net/calibration/mathvision/current/to/videommewithsub/current.md) | linear | 8 | 1.9 pp | 17 |
| [Vals MMLU-Pro → GPQA Diamond](https://benchgap.net/calibration/vals-mmlu-pro/current/to/gpqa-diamond/current.md) | linear | 25 | 1.9 pp | 5 |
| [AIME26 → HMMT Feb 2025](https://benchgap.net/calibration/aime2026/current/to/hmmtfeb2025/current.md) | Hill | 7 | 1.9 pp | 11 |
| [NL2Repo → LiveCodeBench v6](https://benchgap.net/calibration/nl2repo/current/to/livecodebench-v6/current.md) | Michaelis–Menten | 12 | 1.9 pp | 8 |
| [Decision Index Full Score → WildGuardTest (Decision Index)](https://benchgap.net/calibration/decisionindex/current/to/aa-decision-wild-guard-test/current.md) | linear | 11 | 1.9 pp | 2 |
| [HLE → GPQA](https://benchgap.net/calibration/hle/current/to/gpqa/current.md) | Hill | 53 | 1.9 pp | 6 |
| [SkillsBench → BrowseComp](https://benchgap.net/calibration/skills-bench/current/to/browsecomp/current.md) | Hill | 5 | 1.9 pp | 2 |
| [OSWorld 2.0 → ResearchClawBench](https://benchgap.net/calibration/osworld2/current/to/researchclawbench/current.md) | Hill | 5 | 1.9 pp | 2 |
| [ERQA → AA-MMMU-Pro](https://benchgap.net/calibration/erqa/current/to/aa-mmmu-pro/current.md) | linear | 11 | 1.9 pp | 2 |
| [MathVision w/ Python → CharXiv w/o tools](https://benchgap.net/calibration/mathvisionpython/current/to/charxivnotools/current.md) | Michaelis–Menten + offset | 5 | 1.9 pp | 1 |
| [GDPval-AA → DRACO](https://benchgap.net/calibration/gdpvalaanormalized/current/to/draco/current.md) | Michaelis–Menten | 7 | 2.0 pp | 88 |
| [Video-MME (with subtitle) → CountBench](https://benchgap.net/calibration/videommewithsub/current/to/countbench/current.md) | offset logistic | 7 | 2.0 pp | 2 |
| [HMMT Feb 2025 → AIME 2025](https://benchgap.net/calibration/hmmtfeb2025/current/to/aime2025/current.md) | linear | 8 | 2.0 pp | 1 |
| [HMMT Feb 2025 → MMAnswerBench](https://benchgap.net/calibration/hmmtfeb2025/current/to/mmanswerbench/current.md) | Michaelis–Menten + offset | 7 | 2.0 pp | 4 |
| [MMMU → AI2D_TEST](https://benchgap.net/calibration/mmmu/current/to/ai2dtest/current.md) | Hill | 8 | 2.0 pp | 15 |
| [Video-MME (w/o subtitle) → MMMU](https://benchgap.net/calibration/videommenosub/current/to/mmmu/current.md) | Hill | 9 | 2.0 pp | 0 |
| [AI2D_TEST → CountBench](https://benchgap.net/calibration/ai2dtest/current/to/countbench/current.md) | linear | 6 | 2.0 pp | 1 |
| [MAXIFE → MMLU-ProX](https://benchgap.net/calibration/maxife/current/to/mmluprox/current.md) | Michaelis–Menten | 5 | 2.0 pp | 0 |
| [MMMU-Pro w/ Python → AA-MMMU-Pro](https://benchgap.net/calibration/mmmupropython/current/to/aa-mmmu-pro/current.md) | Hill | 10 | 2.0 pp | 0 |
| [PerceptionBench → BabyVision w/ Python](https://benchgap.net/calibration/perceptionbench/current/to/babyvisionpython/current.md) | Michaelis–Menten | 5 | 2.0 pp | 1 |
| [VITA-Bench → QwenClawBench](https://benchgap.net/calibration/vitabench/current/to/qwenclawbench/current.md) | offset logistic | 6 | 2.0 pp | 7 |
| [AA AutomationBench → Hermes Bench](https://benchgap.net/calibration/aa-automation-bench/current/to/hermesbench/current.md) | offset logistic | 6 | 2.0 pp | 0 |
| [SWE Multilingual → INCLUDE](https://benchgap.net/calibration/swe-bench-multilingual/current/to/include/current.md) | linear | 7 | 2.0 pp | 44 |
| [HLE w/o tools → GPQA Diamond](https://benchgap.net/calibration/hlenotools/current/to/gpqa-diamond/current.md) | linear | 36 | 2.0 pp | 3 |
| [Toolathlon → AA Tau3 Banking](https://benchgap.net/calibration/toolathlon/current/to/aa-tau3-banking/current.md) | offset logistic | 6 | 2.0 pp | 11 |
| [ApprenticeBench → AA Harvey LAB v1.0](https://benchgap.net/calibration/apprenticebench/current/to/aa-harvey-lab/current.md) | offset logistic | 5 | 2.0 pp | 2 |
| [CursorBench 4.0 → CursorBench 3.2](https://benchgap.net/calibration/cursor-bench40/current/to/cursor-bench32/current.md) | Michaelis–Menten | 8 | 2.0 pp | 10 |
| [AA Coding Index → CursorBench 4.0](https://benchgap.net/calibration/aa-coding-index/current/to/cursor-bench40/current.md) | inverse Michaelis–Menten | 9 | 2.0 pp | 2 |
| [MMAnswerBench → HMMT Nov 2025](https://benchgap.net/calibration/mmanswerbench/current/to/hmmtnov2025/current.md) | offset logistic | 9 | 2.0 pp | 0 |
| [MathVision → AA-MMMU-Pro](https://benchgap.net/calibration/mathvision/current/to/aa-mmmu-pro/current.md) | Michaelis–Menten | 15 | 2.0 pp | 2 |
| [AIME 2025 → MATH-500](https://benchgap.net/calibration/aime2025/current/to/math-500/current.md) | linear | 17 | 2.0 pp | 24 |
| [When2Call (Decision Index) → HLE-MultiChoice (Decision Index)](https://benchgap.net/calibration/aa-decision-when2-call/current/to/aa-decision-hle-multi-choice/current.md) | Hill | 19 | 2.0 pp | 0 |
| [PostTrainBench v1.1 → PostTrainBench](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1-1/current/to/post-train-bench12/current.md) | linear | 13 | 2.0 pp | 2 |
| [We-Math → RealWorldQA](https://benchgap.net/calibration/wemath/current/to/realworldqa/current.md) | linear | 6 | 2.0 pp | 0 |
| [AA-MMMU-Pro → AI2D_TEST](https://benchgap.net/calibration/aa-mmmu-pro/current/to/ai2dtest/current.md) | offset logistic | 6 | 2.0 pp | 72 |
| [AA-GPQA Diamond → HLE-Verified](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/hleverified/current.md) | offset logistic | 6 | 2.0 pp | 44 |
| [BabyVision → CharXiv w/o tools](https://benchgap.net/calibration/babyvision/current/to/charxivnotools/current.md) | linear | 5 | 2.0 pp | 2 |
| [MMMU-Pro → AI2D_TEST](https://benchgap.net/calibration/mmmu-pro/current/to/ai2dtest/current.md) | Michaelis–Menten | 6 | 2.0 pp | 4 |
| [SimpleVQA → OmniDocBench 1.5](https://benchgap.net/calibration/simplevqa/current/to/omnidocbench15/current.md) | linear | 8 | 2.0 pp | 7 |
| [C-Eval → SuperGPQA](https://benchgap.net/calibration/ceval/current/to/supergpqa/current.md) | Hill | 6 | 2.0 pp | 4 |
| [AA-MMMU-Pro → MMMU-Pro w/ Python](https://benchgap.net/calibration/aa-mmmu-pro/current/to/mmmupropython/current.md) | linear | 10 | 2.1 pp | 0 |
| [HealthBench Professional → AA-GPQA Diamond](https://benchgap.net/calibration/healthbenchprofessional/current/to/aa-gpqa-diamond/current.md) | offset logistic | 7 | 2.1 pp | 12 |
| [OSWorld-Verified → ResearchClawBench](https://benchgap.net/calibration/osworld-verified/current/to/researchclawbench/current.md) | Michaelis–Menten + offset | 10 | 2.1 pp | 12 |
| [GPQA Diamond → Vals GPQA Diamond](https://benchgap.net/calibration/gpqa-diamond/current/to/vals-gpqa-diamond/current.md) | Hill | 25 | 2.1 pp | 42 |
| [Artificial Analysis Decision Index → HLE-MultiChoice (Decision Index)](https://benchgap.net/calibration/aa-decision-index-v2/current/to/aa-decision-hle-multi-choice/current.md) | Hill | 14 | 2.1 pp | 8 |
| [SEC-Bench Pro → DAYJOB Healthcare](https://benchgap.net/calibration/secbenchpro/current/to/dayjob-healthcare/current.md) | Hill | 8 | 2.1 pp | 1 |
| [HLE → MMLU-Redux](https://benchgap.net/calibration/hle/current/to/mmluredux/current.md) | Michaelis–Menten + offset | 9 | 2.1 pp | 7 |
| [MathVision → We-Math](https://benchgap.net/calibration/mathvision/current/to/wemath/current.md) | offset logistic | 6 | 2.1 pp | 19 |
| [ExploitBench Internal Port → DAYJOB](https://benchgap.net/calibration/exploitbenchinternalport/current/to/dayjob/current.md) | Hill | 5 | 2.1 pp | 0 |
| [ERQA → CharXiv w/o tools](https://benchgap.net/calibration/erqa/current/to/charxivnotools/current.md) | Michaelis–Menten | 5 | 2.1 pp | 1 |
| [ARC-AGI-3 Provider Adapter → HealthBench Professional](https://benchgap.net/calibration/arc-agi-3-provider-adapter/current/to/healthbenchprofessional/current.md) | offset logistic | 7 | 2.1 pp | 6 |
| [VideoMMMU → OCRBench V2](https://benchgap.net/calibration/videommmu/current/to/ocrbenchv2/current.md) | Hill | 6 | 2.1 pp | 10 |
| [Video-MME (with subtitle) → DynaMath](https://benchgap.net/calibration/videommewithsub/current/to/dynamath/current.md) | linear | 7 | 2.1 pp | 1 |
| [FrontierCode 1.1 Extended → CursorBench 3.2](https://benchgap.net/calibration/frontiercode11extended/current/to/cursor-bench32/current.md) | Michaelis–Menten + offset | 5 | 2.1 pp | 0 |
| [ERQA → We-Math](https://benchgap.net/calibration/erqa/current/to/wemath/current.md) | Michaelis–Menten | 5 | 2.1 pp | 7 |
| [MMLU-Pro → MMLU](https://benchgap.net/calibration/mmlu-pro/current/to/mmlu/current.md) | Michaelis–Menten + offset | 19 | 2.1 pp | 27 |
| [Video-MME (w/o subtitle) → AA-MMMU-Pro](https://benchgap.net/calibration/videommenosub/current/to/aa-mmmu-pro/current.md) | Hill | 6 | 2.1 pp | 0 |
| [OSWorld-Verified → WideResearch](https://benchgap.net/calibration/osworld-verified/current/to/wideresearch/current.md) | offset logistic | 5 | 2.1 pp | 17 |
| [Chess Puzzles → FailureSensorIQ](https://benchgap.net/calibration/chesspuzzles/current/to/failuresensoriq/current.md) | linear | 9 | 2.1 pp | 24 |
| [Decision Index Tools → WildGuardTest (Decision Index)](https://benchgap.net/calibration/jevdecisionindex03tools/current/to/aa-decision-wild-guard-test/current.md) | linear | 11 | 2.1 pp | 0 |
| [AA-HLE → HealthBench Professional (raw, Artificial Analysis)](https://benchgap.net/calibration/aa-hle/current/to/aa-health-bench-professional-raw/current.md) | Hill | 6 | 2.1 pp | 12 |
| [Bug Hunt Bench → SWE-Marathon](https://benchgap.net/calibration/bug-hunt-bench/current/to/swemarathon/current.md) | linear | 5 | 2.1 pp | 25 |
| [Decision Index Knowledge → HLE-MultiChoice (Decision Index)](https://benchgap.net/calibration/jevdecisionindex03knowledge/current/to/aa-decision-hle-multi-choice/current.md) | offset logistic | 11 | 2.1 pp | 82 |
| [MMMLU → GPQA Diamond](https://benchgap.net/calibration/mmmlu/current/to/gpqa-diamond/current.md) | Hill | 5 | 2.1 pp | 0 |
| [ERQA → CountBench](https://benchgap.net/calibration/erqa/current/to/countbench/current.md) | linear | 6 | 2.2 pp | 6 |
| [OTelBench → Mystery Game Puzzles](https://benchgap.net/calibration/otelbench/current/to/mysterygamepuzzles/current.md) | Michaelis–Menten | 5 | 2.2 pp | 8 |
| [VideoMMMU → MMMU-Pro](https://benchgap.net/calibration/videommmu/current/to/mmmu-pro/current.md) | Hill | 16 | 2.2 pp | 0 |
| [Decision Index Retrieval → WildGuardTest (Decision Index)](https://benchgap.net/calibration/jevdecisionindex03retrieval/current/to/aa-decision-wild-guard-test/current.md) | offset logistic | 11 | 2.2 pp | 10 |
| [BBH → MMLU](https://benchgap.net/calibration/bbh/current/to/mmlu/current.md) | linear | 6 | 2.2 pp | 0 |
| [GDPval-AA → ResearchClawBench](https://benchgap.net/calibration/gdpvalaanormalized/current/to/researchclawbench/current.md) | linear | 14 | 2.2 pp | 47 |
| [MathVision → OmniDocBench 1.5](https://benchgap.net/calibration/mathvision/current/to/omnidocbench15/current.md) | linear | 8 | 2.2 pp | 13 |
| [GraphWalks BFS 256K–1M → ARC-AGI-2](https://benchgap.net/calibration/graphwalks-bfs256k1m/current/to/arc-agi-2/current.md) | offset logistic | 5 | 2.2 pp | 0 |
| [ARC-AGI-1 → Vals MMLU-Pro](https://benchgap.net/calibration/arcagi1/current/to/vals-mmlu-pro/current.md) | Michaelis–Menten | 28 | 2.2 pp | 0 |
| [HealthBench Professional → HealthBench (length-adjusted)](https://benchgap.net/calibration/healthbenchprofessional/current/to/healthbenchlengthadjusted/current.md) | inverse Michaelis–Menten | 8 | 2.2 pp | 6 |
| [LABBench2 → Vals GPQA Diamond](https://benchgap.net/calibration/labbench2/current/to/vals-gpqa-diamond/current.md) | Michaelis–Menten + offset | 6 | 2.2 pp | 0 |
| [PostTrainBench → PostTrainBench v1.1](https://benchgap.net/calibration/post-train-bench12/current/to/posttrainbench%23posttrainbench-v1-1/current.md) | linear | 13 | 2.2 pp | 6 |
| [MMMU → MMMU-Pro](https://benchgap.net/calibration/mmmu/current/to/mmmu-pro/current.md) | inverse Michaelis–Menten | 19 | 2.2 pp | 9 |
| [Vals LiveCodeBench → LiveCodeBench v6](https://benchgap.net/calibration/valslivecodebench/current/to/livecodebench-v6/current.md) | Michaelis–Menten + offset | 7 | 2.2 pp | 56 |
| [When2Call (Decision Index) → Artificial Analysis Decision Index v1.0](https://benchgap.net/calibration/aa-decision-when2-call/current/to/aa-decision-index/current.md) | Hill | 19 | 2.2 pp | 0 |
| [cursorBench41 → AA-SciCode](https://benchgap.net/calibration/cursor-bench41/current/to/aa-sci-code/current.md) | Hill | 9 | 2.2 pp | 3 |
| [MLVU (M-Avg) → DynaMath](https://benchgap.net/calibration/mlvuavg/current/to/dynamath/current.md) | Michaelis–Menten | 7 | 2.2 pp | 0 |
| [ZeroBench → MathVision](https://benchgap.net/calibration/zerobench/current/to/mathvision/current.md) | Hill | 6 | 2.2 pp | 2 |
| [MMMU-Pro → MLVU (M-Avg)](https://benchgap.net/calibration/mmmu-pro/current/to/mlvuavg/current.md) | linear | 10 | 2.2 pp | 26 |
| [Artificial Analysis Decision Index → WildGuardTest (Decision Index)](https://benchgap.net/calibration/aa-decision-index-v2/current/to/aa-decision-wild-guard-test/current.md) | Michaelis–Menten | 14 | 2.2 pp | 5 |
| [CyberGym → Toolathlon-Verified](https://benchgap.net/calibration/cybergym/current/to/toolathlonverified/current.md) | Michaelis–Menten | 7 | 2.2 pp | 9 |
| [FrontierCode 1.1 Main → SWE-bench Verified](https://benchgap.net/calibration/frontiercode/current/to/swe-bench-verified/current.md) | linear | 7 | 2.2 pp | 0 |
| [Decision Index Retrieval → HLE-MultiChoice (Decision Index)](https://benchgap.net/calibration/jevdecisionindex03retrieval/current/to/aa-decision-hle-multi-choice/current.md) | offset logistic | 11 | 2.2 pp | 8 |
| [Gert Labs → ResearchClawBench](https://benchgap.net/calibration/gertlabs/current/to/researchclawbench/current.md) | linear | 16 | 2.2 pp | 11 |
| [NOVA-63 → MAXIFE](https://benchgap.net/calibration/nova63/current/to/maxife/current.md) | linear | 5 | 2.2 pp | 2 |
| [GPQA → BioBasicsBench](https://benchgap.net/calibration/gpqa/current/to/biobasicsbench/current.md) | inverse Michaelis–Menten | 5 | 2.2 pp | 32 |
| [Toolathlon → ResearchClawBench](https://benchgap.net/calibration/toolathlon/current/to/researchclawbench/current.md) | Michaelis–Menten | 9 | 2.3 pp | 0 |
| [CursorBench 3.2 → VulcanBench Frontier v4](https://benchgap.net/calibration/cursor-bench32/current/to/vulcanbench/current.md) | Hill | 5 | 2.3 pp | 6 |
| [AI2D_TEST → MLVU (M-Avg)](https://benchgap.net/calibration/ai2dtest/current/to/mlvuavg/current.md) | Michaelis–Menten | 6 | 2.3 pp | 1 |
| [MathVision → RealWorldQA](https://benchgap.net/calibration/mathvision/current/to/realworldqa/current.md) | offset logistic | 11 | 2.3 pp | 1 |
| [AA Agentic Index → Toolathlon-Verified](https://benchgap.net/calibration/aa-agentic-index/current/to/toolathlonverified/current.md) | Michaelis–Menten | 7 | 2.3 pp | 46 |
| [MMMU → VideoMMMU](https://benchgap.net/calibration/mmmu/current/to/videommmu/current.md) | Hill | 12 | 2.3 pp | 15 |
| [Decision Index Retrieval → Artificial Analysis Decision Index](https://benchgap.net/calibration/jevdecisionindex03retrieval/current/to/aa-decision-index-v2/current.md) | linear | 11 | 2.3 pp | 0 |
| [Video-MME (w/o subtitle) → MMMU-Pro](https://benchgap.net/calibration/videommenosub/current/to/mmmu-pro/current.md) | linear | 8 | 2.3 pp | 0 |
| [OfficeQA Pro → AA-MMMU-Pro](https://benchgap.net/calibration/officeqapro/current/to/aa-mmmu-pro/current.md) | offset logistic | 9 | 2.3 pp | 7 |
| [OmniDocBench 1.5 → MLVU (M-Avg)](https://benchgap.net/calibration/omnidocbench15/current/to/mlvuavg/current.md) | Michaelis–Menten + offset | 8 | 2.3 pp | 0 |
| [AA-AnalystAgent → LifeVal](https://benchgap.net/calibration/aa-analyst-agent/current/to/lifeval/current.md) | Michaelis–Menten | 5 | 2.3 pp | 15 |
| [Decision Index Full Score → HLE-MultiChoice (Decision Index)](https://benchgap.net/calibration/decisionindex/current/to/aa-decision-hle-multi-choice/current.md) | Michaelis–Menten | 11 | 2.3 pp | 0 |
| [ARC-AGI-3 Provider Adapter → HealthBench Professional (raw)](https://benchgap.net/calibration/arc-agi-3-provider-adapter/current/to/healthbenchprofessionalraw/current.md) | linear | 6 | 2.3 pp | 7 |
| [NL2Repo → SWE-bench Verified](https://benchgap.net/calibration/nl2repo/current/to/swe-bench-verified/current.md) | Hill | 19 | 2.3 pp | 17 |
| [Decision Index Knowledge → WildGuardTest (Decision Index)](https://benchgap.net/calibration/jevdecisionindex03knowledge/current/to/aa-decision-wild-guard-test/current.md) | offset logistic | 11 | 2.3 pp | 2 |
| [C-Eval → HLE](https://benchgap.net/calibration/ceval/current/to/hle/current.md) | linear | 6 | 2.3 pp | 4 |
| [DeepSearchQA → AA EnterpriseOps-Gym](https://benchgap.net/calibration/deepsearchqa/current/to/aa-enterprise-ops-gym/current.md) | Michaelis–Menten | 5 | 2.3 pp | 12 |
| [FrontierCode 2.0 → AA Coding Index](https://benchgap.net/calibration/frontiercode-2/current/to/aa-coding-index/current.md) | Michaelis–Menten | 7 | 2.3 pp | 15 |
| [GPQA → Vals MMLU-Pro](https://benchgap.net/calibration/gpqa/current/to/vals-mmlu-pro/current.md) | Michaelis–Menten | 40 | 2.3 pp | 70 |
| [ARC-AGI-2 → Vals GPQA Diamond](https://benchgap.net/calibration/arc-agi-2/current/to/vals-gpqa-diamond/current.md) | Michaelis–Menten + offset | 31 | 2.3 pp | 5 |
| [HLE → GPQA Diamond](https://benchgap.net/calibration/hle/current/to/gpqa-diamond/current.md) | Michaelis–Menten | 42 | 2.3 pp | 4 |
| [HealthBench Professional (raw) → HealthBench (length-adjusted)](https://benchgap.net/calibration/healthbenchprofessionalraw/current/to/healthbenchlengthadjusted/current.md) | linear | 8 | 2.3 pp | 0 |
| [BrowseComp → ResearchClawBench](https://benchgap.net/calibration/browsecomp/current/to/researchclawbench/current.md) | Michaelis–Menten | 10 | 2.3 pp | 7 |
| [ApprenticeBench → OSWorld-Verified](https://benchgap.net/calibration/apprenticebench/current/to/osworld-verified/current.md) | linear | 7 | 2.3 pp | 0 |
| [ExploitBench Internal Port → SpireClimb](https://benchgap.net/calibration/exploitbenchinternalport/current/to/spireclimb/current.md) | Michaelis–Menten | 6 | 2.4 pp | 3 |
| [Decision Index Tools → HLE-MultiChoice (Decision Index)](https://benchgap.net/calibration/jevdecisionindex03tools/current/to/aa-decision-hle-multi-choice/current.md) | linear | 11 | 2.4 pp | 0 |
| [Hermes Index → Toolathlon-Verified](https://benchgap.net/calibration/hermesindex/current/to/toolathlonverified/current.md) | linear | 8 | 2.4 pp | 0 |
| [MMLU → Vals MMLU-Pro](https://benchgap.net/calibration/mmlu/current/to/vals-mmlu-pro/current.md) | linear | 8 | 2.4 pp | 5 |
| [AA-HLE → LABBench2](https://benchgap.net/calibration/aa-hle/current/to/labbench2/current.md) | linear | 7 | 2.4 pp | 25 |
| [APEX-Agents-AA → ResearchClawBench](https://benchgap.net/calibration/apexagentsaa/current/to/researchclawbench/current.md) | linear | 9 | 2.4 pp | 3 |
| [MMMU-Pro → MedXpertQA (MM)](https://benchgap.net/calibration/mmmu-pro/current/to/medxpertqamm/current.md) | Hill | 8 | 2.4 pp | 44 |
| [Vals SWE-bench → FrontierCode 1.1 Extended](https://benchgap.net/calibration/valsswebench/current/to/frontiercode11extended/current.md) | Hill | 5 | 2.4 pp | 3 |
| [CWE-Bench-AA → AA Tau3 Banking](https://benchgap.net/calibration/aa-cwe-bench/current/to/aa-tau3-banking/current.md) | offset logistic | 5 | 2.4 pp | 16 |
| [Decision Index Language → HLE-MultiChoice (Decision Index)](https://benchgap.net/calibration/jevdecisionindex03language/current/to/aa-decision-hle-multi-choice/current.md) | Michaelis–Menten | 11 | 2.4 pp | 0 |
| [MMMU-Pro → VideoMMMU](https://benchgap.net/calibration/mmmu-pro/current/to/videommmu/current.md) | Hill | 16 | 2.4 pp | 20 |
| [Decision Index Arts → HLE-MultiChoice (Decision Index)](https://benchgap.net/calibration/jevdecisionindex03arts/current/to/aa-decision-hle-multi-choice/current.md) | Michaelis–Menten + offset | 11 | 2.4 pp | 0 |
| [ARC-AGI-1 → Vals GPQA Diamond](https://benchgap.net/calibration/arcagi1/current/to/vals-gpqa-diamond/current.md) | Michaelis–Menten | 28 | 2.4 pp | 0 |
| [MMMU → MLVU (M-Avg)](https://benchgap.net/calibration/mmmu/current/to/mlvuavg/current.md) | Hill | 8 | 2.4 pp | 11 |
| [FrontierSWE v2 → PostTrainBench](https://benchgap.net/calibration/frontierswev2/current/to/post-train-bench12/current.md) | offset logistic | 10 | 2.4 pp | 14 |
| [MMAnswerBench → HMMT Feb 2025](https://benchgap.net/calibration/mmanswerbench/current/to/hmmtfeb2025/current.md) | Michaelis–Menten | 7 | 2.4 pp | 0 |
| [SuperGPQA → GPQA](https://benchgap.net/calibration/supergpqa/current/to/gpqa/current.md) | Hill | 16 | 2.4 pp | 3 |
| [AA-HLE → GPQA Diamond](https://benchgap.net/calibration/aa-hle/current/to/gpqa-diamond/current.md) | offset logistic | 50 | 2.4 pp | 61 |
| [AA-SciCode → OpenHarmony Bench v1.0](https://benchgap.net/calibration/aa-sci-code/current/to/openharmonybench%23openharmonybench-v1-0/current.md) | offset logistic | 11 | 2.4 pp | 92 |
| [MMLU → AA-Omniscience Accuracy](https://benchgap.net/calibration/mmlu/current/to/omniscienceaccuracy/current.md) | offset logistic | 10 | 2.4 pp | 21 |
| [CharXiv w/o tools → CharXiv](https://benchgap.net/calibration/charxivnotools/current/to/charxiv/current.md) | Michaelis–Menten | 15 | 2.4 pp | 2 |
| [Video-MME (with subtitle) → MLVU (M-Avg)](https://benchgap.net/calibration/videommewithsub/current/to/mlvuavg/current.md) | Michaelis–Menten | 10 | 2.4 pp | 0 |
| [Video-MME (with subtitle) → AA-MMMU-Pro](https://benchgap.net/calibration/videommewithsub/current/to/aa-mmmu-pro/current.md) | Michaelis–Menten | 8 | 2.4 pp | 0 |
| [When2Call (Decision Index) → Decision Index Arts](https://benchgap.net/calibration/aa-decision-when2-call/current/to/jevdecisionindex03arts/current.md) | linear | 11 | 2.4 pp | 8 |
| [FrontierMath v2 (Tiers 1-3) → MMAnswerBench](https://benchgap.net/calibration/frontiermathv2tiers13/current/to/mmanswerbench/current.md) | linear | 7 | 2.4 pp | 62 |
| [CWE-bench v1 → Hermes Bench](https://benchgap.net/calibration/cwe-bench%23cwe-bench-v1/current/to/hermesbench/current.md) | linear | 5 | 2.4 pp | 4 |
| [RealWorldQA → AA-MMMU-Pro](https://benchgap.net/calibration/realworldqa/current/to/aa-mmmu-pro/current.md) | Michaelis–Menten | 9 | 2.4 pp | 1 |
| [CharXiv → CharXiv w/o tools](https://benchgap.net/calibration/charxiv/current/to/charxivnotools/current.md) | offset logistic | 15 | 2.4 pp | 5 |
| [AI2D_TEST → MMMU-Pro](https://benchgap.net/calibration/ai2dtest/current/to/mmmu-pro/current.md) | inverse Michaelis–Menten | 6 | 2.4 pp | 0 |
| [ARC-AGI-3 → HLE](https://benchgap.net/calibration/arcagi3/current/to/hle/current.md) | linear | 8 | 2.4 pp | 0 |
| [MMLU → BBH](https://benchgap.net/calibration/mmlu/current/to/bbh/current.md) | linear | 6 | 2.4 pp | 29 |
| [HLE → Vals MMLU-Pro](https://benchgap.net/calibration/hle/current/to/vals-mmlu-pro/current.md) | linear | 32 | 2.4 pp | 6 |
| [GMMLU → SWE Multilingual](https://benchgap.net/calibration/gmmlu/current/to/swe-bench-multilingual/current.md) | Hill | 5 | 2.4 pp | 0 |
| [AI2D_TEST → DynaMath](https://benchgap.net/calibration/ai2dtest/current/to/dynamath/current.md) | Michaelis–Menten | 5 | 2.4 pp | 1 |
| [LVBench → CharXiv w/o tools](https://benchgap.net/calibration/lvbench/current/to/charxivnotools/current.md) | Michaelis–Menten | 6 | 2.5 pp | 1 |
| [DeepSWE → AA Coding Index](https://benchgap.net/calibration/deepswe/current/to/aa-coding-index/current.md) | linear | 20 | 2.5 pp | 28 |
| [DeepsecBench-AA → Toolathlon-Verified](https://benchgap.net/calibration/aa-deepsec-bench/current/to/toolathlonverified/current.md) | Michaelis–Menten | 8 | 2.5 pp | 1 |
| [ERQA → OmniDocBench 1.5](https://benchgap.net/calibration/erqa/current/to/omnidocbench15/current.md) | linear | 7 | 2.5 pp | 2 |
| [SWE-bench Multimodal → OfficeQA Pro](https://benchgap.net/calibration/swe-bench-multimodal/current/to/officeqapro/current.md) | Hill | 6 | 2.5 pp | 2 |
| [Artificial Analysis Decision Index → Decision Index Full Score](https://benchgap.net/calibration/aa-decision-index-v2/current/to/decisionindex/current.md) | inverse Michaelis–Menten | 11 | 2.5 pp | 11 |
| [CursorBench 4.0 → AA-SciCode](https://benchgap.net/calibration/cursor-bench40/current/to/aa-sci-code/current.md) | Hill | 17 | 2.5 pp | 1 |
| [Terminal-Bench 2.1 (Vals AI) → AA Terminal-Bench 2.1](https://benchgap.net/calibration/vals-terminal-bench21/current/to/aa-terminal-bench21/current.md) | Hill | 18 | 2.5 pp | 48 |
| [cursorBench41 → VulcanBench Frontier v4](https://benchgap.net/calibration/cursor-bench41/current/to/vulcanbench/current.md) | Hill | 6 | 2.5 pp | 1 |
| [GPQA Diamond → ARC-AGI-1](https://benchgap.net/calibration/gpqa-diamond/current/to/arcagi1/current.md) | Hill | 15 | 2.5 pp | 47 |
| [PerceptionBench → CharXiv w/o tools](https://benchgap.net/calibration/perceptionbench/current/to/charxivnotools/current.md) | linear | 6 | 2.5 pp | 0 |
| [BabyVision → MMMU-Pro](https://benchgap.net/calibration/babyvision/current/to/mmmu-pro/current.md) | linear | 6 | 2.5 pp | 4 |
| [PostTrainBench → AA Coding Index](https://benchgap.net/calibration/post-train-bench12/current/to/aa-coding-index/current.md) | Michaelis–Menten | 12 | 2.5 pp | 0 |
| [Video-MME (w/o subtitle) → AI2D_TEST](https://benchgap.net/calibration/videommenosub/current/to/ai2dtest/current.md) | Michaelis–Menten + offset | 7 | 2.5 pp | 0 |
| [ARC-AGI-3 → LABBench2](https://benchgap.net/calibration/arcagi3/current/to/labbench2/current.md) | linear | 5 | 2.5 pp | 2 |
| [BabyVision w/ Python → PerceptionBench](https://benchgap.net/calibration/babyvisionpython/current/to/perceptionbench/current.md) | inverse Michaelis–Menten | 5 | 2.5 pp | 4 |
| [Artificial Analysis Decision Index v1.0 → WildGuardTest (Decision Index)](https://benchgap.net/calibration/aa-decision-index/current/to/aa-decision-wild-guard-test/current.md) | linear | 19 | 2.5 pp | 0 |
| [Chess Puzzles → Cookie-Bench](https://benchgap.net/calibration/chesspuzzles/current/to/cookie-bench/current.md) | linear | 5 | 2.5 pp | 19 |
| [ARC-AGI-1 → HealthBench (length-adjusted)](https://benchgap.net/calibration/arcagi1/current/to/healthbenchlengthadjusted/current.md) | Michaelis–Menten | 6 | 2.5 pp | 15 |
| [OmniDocBench 1.5 → CountBench](https://benchgap.net/calibration/omnidocbench15/current/to/countbench/current.md) | Hill | 5 | 2.5 pp | 0 |
| [Decision Index Tools → Artificial Analysis Decision Index](https://benchgap.net/calibration/jevdecisionindex03tools/current/to/aa-decision-index-v2/current.md) | Hill | 11 | 2.5 pp | 0 |
| [AA-Omniscience Accuracy → HealthBench Professional (raw, Artificial Analysis)](https://benchgap.net/calibration/omniscienceaccuracy/current/to/aa-health-bench-professional-raw/current.md) | Hill | 6 | 2.5 pp | 24 |
| [HLE w/o tools → MMLU-Pro](https://benchgap.net/calibration/hlenotools/current/to/mmlu-pro/current.md) | Michaelis–Menten | 7 | 2.5 pp | 49 |
| [AA-MMMU-Pro → CharXiv w/o tools](https://benchgap.net/calibration/aa-mmmu-pro/current/to/charxivnotools/current.md) | Michaelis–Menten + offset | 11 | 2.5 pp | 30 |
| [MCP-Tasks → WideResearch](https://benchgap.net/calibration/mcptasks/current/to/wideresearch/current.md) | linear | 5 | 2.5 pp | 0 |
| [MMMU-Pro → MMMU-Pro w/ Python](https://benchgap.net/calibration/mmmu-pro/current/to/mmmupropython/current.md) | Michaelis–Menten | 10 | 2.6 pp | 18 |
| [cursorBench41 → CursorBench 4.0](https://benchgap.net/calibration/cursor-bench41/current/to/cursor-bench40/current.md) | linear | 11 | 2.6 pp | 1 |
| [SEC-Bench Pro → DAYJOB](https://benchgap.net/calibration/secbenchpro/current/to/dayjob/current.md) | Hill | 8 | 2.6 pp | 1 |
| [MazeBench → Toolathlon-Verified](https://benchgap.net/calibration/mazebench/current/to/toolathlonverified/current.md) | linear | 10 | 2.6 pp | 5 |
| [Artificial Analysis Intelligence Index → GPQA Diamond](https://benchgap.net/calibration/artificialanalysis/current/to/gpqa-diamond/current.md) | Michaelis–Menten | 42 | 2.6 pp | 8 |
| [Artificial Analysis Decision Index → Decision Index Retrieval](https://benchgap.net/calibration/aa-decision-index-v2/current/to/jevdecisionindex03retrieval/current.md) | linear | 11 | 2.6 pp | 11 |
| [HealthBench (raw) → HealthBench (length-adjusted)](https://benchgap.net/calibration/healthbench/current/to/healthbenchlengthadjusted/current.md) | Michaelis–Menten | 8 | 2.6 pp | 0 |
| [MMAnswerBench → HMMT Feb 2026](https://benchgap.net/calibration/mmanswerbench/current/to/hmmtfeb2026/current.md) | Michaelis–Menten | 10 | 2.6 pp | 0 |
| [Vals SWE-bench → AA LiveCodeBench](https://benchgap.net/calibration/valsswebench/current/to/aa-live-code-bench/current.md) | Hill | 9 | 2.6 pp | 63 |
| [FrontierSWE v2 → CursorBench 4.0](https://benchgap.net/calibration/frontierswev2/current/to/cursor-bench40/current.md) | linear | 15 | 2.6 pp | 3 |
| [AA-LCR → MMLongBench-Doc](https://benchgap.net/calibration/lcr/current/to/mmlongbenchdoc/current.md) | Michaelis–Menten + offset | 6 | 2.6 pp | 191 |
| [WildGuardTest (Decision Index) → HLE-MultiChoice (Decision Index)](https://benchgap.net/calibration/aa-decision-wild-guard-test/current/to/aa-decision-hle-multi-choice/current.md) | linear | 19 | 2.6 pp | 0 |
| [AI2D_TEST → MathVision](https://benchgap.net/calibration/ai2dtest/current/to/mathvision/current.md) | inverse Michaelis–Menten | 5 | 2.6 pp | 3 |
| [HLE w/o tools → HealthBench Professional (raw, Artificial Analysis)](https://benchgap.net/calibration/hlenotools/current/to/aa-health-bench-professional-raw/current.md) | Michaelis–Menten | 5 | 2.6 pp | 11 |
| [MCP Atlas → QwenClawBench](https://benchgap.net/calibration/mcpatlas/current/to/qwenclawbench/current.md) | Hill | 8 | 2.6 pp | 44 |
| [Decision Index Full Score → Decision Index Knowledge](https://benchgap.net/calibration/decisionindex/current/to/jevdecisionindex03knowledge/current.md) | inverse Michaelis–Menten | 110 | 2.6 pp | 0 |
| [AA-HLE → MMMLU](https://benchgap.net/calibration/aa-hle/current/to/mmmlu/current.md) | linear | 9 | 2.6 pp | 10 |
| [JobBench → ResearchClawBench](https://benchgap.net/calibration/jobbench/current/to/researchclawbench/current.md) | linear | 5 | 2.6 pp | 4 |
| [Terminal-Bench 2.1 → AA Terminal-Bench 2.1](https://benchgap.net/calibration/terminal-bench%23terminal-bench-2-1/current/to/aa-terminal-bench21/current.md) | offset logistic | 14 | 2.6 pp | 40 |
| [CursorBench 4.0 → cursorBench41](https://benchgap.net/calibration/cursor-bench40/current/to/cursor-bench41/current.md) | linear | 11 | 2.6 pp | 2 |
| [HLE-MultiChoice (Decision Index) → WildGuardTest (Decision Index)](https://benchgap.net/calibration/aa-decision-hle-multi-choice/current/to/aa-decision-wild-guard-test/current.md) | Michaelis–Menten + offset | 19 | 2.6 pp | 0 |
| [Chartography (tools) → OfficeQA](https://benchgap.net/calibration/chartographywithtools/current/to/officeqa/current.md) | offset logistic | 5 | 2.6 pp | 2 |
| [MedXpertQA (MM) → AA-MMMU-Pro](https://benchgap.net/calibration/medxpertqamm/current/to/aa-mmmu-pro/current.md) | linear | 6 | 2.6 pp | 0 |
| [Decision Index Language → Artificial Analysis Decision Index](https://benchgap.net/calibration/jevdecisionindex03language/current/to/aa-decision-index-v2/current.md) | Hill | 11 | 2.6 pp | 3 |
| [AIME 2025 → AIME 2024](https://benchgap.net/calibration/aime2025/current/to/aime2024/current.md) | Michaelis–Menten | 17 | 2.6 pp | 32 |
| [FrontierCode 1.1 Extended → AA Coding Index](https://benchgap.net/calibration/frontiercode11extended/current/to/aa-coding-index/current.md) | offset logistic | 6 | 2.6 pp | 0 |
| [When2Call (Decision Index) → WildGuardTest (Decision Index)](https://benchgap.net/calibration/aa-decision-when2-call/current/to/aa-decision-wild-guard-test/current.md) | Hill | 19 | 2.7 pp | 0 |
| [AA-GPQA Diamond → GPQA Diamond](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/gpqa-diamond/current.md) | Hill | 41 | 2.7 pp | 0 |
| [DRACO → BrowseComp](https://benchgap.net/calibration/draco/current/to/browsecomp/current.md) | Hill | 5 | 2.7 pp | 2 |
| [Chartography (no tools) → Chartography (tools)](https://benchgap.net/calibration/chartography/current/to/chartographywithtools/current.md) | Michaelis–Menten | 7 | 2.7 pp | 5 |
| [SWE-bench Multimodal → OfficeQA](https://benchgap.net/calibration/swe-bench-multimodal/current/to/officeqa/current.md) | offset logistic | 5 | 2.7 pp | 1 |
| [ExploitBench → FACTS Multimodal](https://benchgap.net/calibration/exploitbench/current/to/facts-multimodal/current.md) | linear | 6 | 2.7 pp | 12 |
| [Vals GPQA Diamond → MMLU](https://benchgap.net/calibration/vals-gpqa-diamond/current/to/mmlu/current.md) | linear | 8 | 2.7 pp | 0 |
| [ExploitBench → FrontierPhysics](https://benchgap.net/calibration/exploitbench/current/to/frontierphysics/current.md) | inverse Michaelis–Menten | 12 | 2.7 pp | 6 |
| [BBH → MMLU-Pro](https://benchgap.net/calibration/bbh/current/to/mmlu-pro/current.md) | inverse Michaelis–Menten | 8 | 2.7 pp | 0 |
| [Protocols (troubleshooting) → SingleCellBench](https://benchgap.net/calibration/protocolstroubleshooting/current/to/singlecellbench/current.md) | Hill | 5 | 2.7 pp | 0 |
| [SWE-bench Verified → React Native Evals](https://benchgap.net/calibration/swe-bench-verified/current/to/reactnativeevals/current.md) | linear | 7 | 2.7 pp | 0 |
| [CursorBench 3.2 → AA-SciCode](https://benchgap.net/calibration/cursor-bench32/current/to/aa-sci-code/current.md) | linear | 17 | 2.7 pp | 0 |
| [Furniture Assembly → DAYJOB](https://benchgap.net/calibration/furnitureassembly/current/to/dayjob/current.md) | offset logistic | 16 | 2.7 pp | 19 |
| [OSWorld 2.1 Offline → AA-AnalystAgent](https://benchgap.net/calibration/osworld-2-1-offline/current/to/aa-analyst-agent/current.md) | linear | 6 | 2.7 pp | 5 |
| [APEX-Agents-AA → AA Tau3 Banking](https://benchgap.net/calibration/apexagentsaa/current/to/aa-tau3-banking/current.md) | offset logistic | 6 | 2.7 pp | 10 |
| [ARC-AGI-2 → AA-GPQA Diamond](https://benchgap.net/calibration/arc-agi-2/current/to/aa-gpqa-diamond/current.md) | linear | 30 | 2.7 pp | 9 |
| [MLVU (M-Avg) → AI2D_TEST](https://benchgap.net/calibration/mlvuavg/current/to/ai2dtest/current.md) | Michaelis–Menten | 6 | 2.7 pp | 0 |
| [HLE w/o tools → AA-GPQA Diamond](https://benchgap.net/calibration/hlenotools/current/to/aa-gpqa-diamond/current.md) | Michaelis–Menten + offset | 25 | 2.7 pp | 23 |
| [SimpleQA → AA-Omniscience Accuracy](https://benchgap.net/calibration/simpleqa/current/to/omniscienceaccuracy/current.md) | Michaelis–Menten | 7 | 2.7 pp | 4 |
| [Terminal-Bench 3.0 → AA Terminal-Bench 2.1](https://benchgap.net/calibration/terminal-bench%23terminal-bench-3-0/current/to/aa-terminal-bench21/current.md) | Michaelis–Menten | 9 | 2.7 pp | 0 |
| [We-Math → ERQA](https://benchgap.net/calibration/wemath/current/to/erqa/current.md) | inverse Michaelis–Menten | 5 | 2.7 pp | 1 |
| [SWE-bench Pro → LiveCodeBench v6](https://benchgap.net/calibration/swe-bench-pro/current/to/livecodebench-v6/current.md) | Michaelis–Menten + offset | 22 | 2.7 pp | 21 |
| [GDPval-AA → BFCL v4](https://benchgap.net/calibration/gdpvalaanormalized/current/to/bfcl-v4/current.md) | Michaelis–Menten + offset | 9 | 2.7 pp | 118 |
| [AA MMLU-Pro → CritPt](https://benchgap.net/calibration/aa-mmlu-pro/current/to/critpt/current.md) | offset logistic | 12 | 2.7 pp | 1 |
| [ARC-AGI-1 → AA-GPQA Diamond](https://benchgap.net/calibration/arcagi1/current/to/aa-gpqa-diamond/current.md) | linear | 28 | 2.7 pp | 0 |
| [GDP.pdf → Hermes Index](https://benchgap.net/calibration/aa-gdp-pdf/current/to/hermesindex/current.md) | offset logistic | 6 | 2.7 pp | 17 |
| [AA-GPQA Diamond → AA-HLE](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/aa-hle/current.md) | offset logistic | 190 | 2.7 pp | 3 |
| [MMMU-Pro → MMVU](https://benchgap.net/calibration/mmmu-pro/current/to/mmvu/current.md) | Michaelis–Menten | 7 | 2.7 pp | 52 |
| [AA-MMMU-Pro → MMMU](https://benchgap.net/calibration/aa-mmmu-pro/current/to/mmmu/current.md) | Michaelis–Menten | 15 | 2.7 pp | 65 |
| [MILU → SWE Multilingual](https://benchgap.net/calibration/milu/current/to/swe-bench-multilingual/current.md) | Hill | 5 | 2.7 pp | 0 |
| [ParseBench-1 → MMMU-Pro w/ Python](https://benchgap.net/calibration/parse-bench-llama-index/current/to/mmmupropython/current.md) | Michaelis–Menten | 6 | 2.7 pp | 10 |
| [Decision Index Arts → Artificial Analysis Decision Index](https://benchgap.net/calibration/jevdecisionindex03arts/current/to/aa-decision-index-v2/current.md) | linear | 11 | 2.7 pp | 2 |
| [ZeroBench → PerceptionBench](https://benchgap.net/calibration/zerobench/current/to/perceptionbench/current.md) | offset logistic | 5 | 2.7 pp | 6 |
| [NL2Repo → OpenHarmony Bench v1.0](https://benchgap.net/calibration/nl2repo/current/to/openharmonybench%23openharmonybench-v1-0/current.md) | Hill | 10 | 2.7 pp | 18 |
| [PerceptionBench → MathVision](https://benchgap.net/calibration/perceptionbench/current/to/mathvision/current.md) | Hill | 5 | 2.7 pp | 1 |
| [OTelBench → FailureSensorIQ](https://benchgap.net/calibration/otelbench/current/to/failuresensoriq/current.md) | Michaelis–Menten | 5 | 2.8 pp | 1 |
| [HMMT Feb 2026 → MMAnswerBench](https://benchgap.net/calibration/hmmtfeb2026/current/to/mmanswerbench/current.md) | Michaelis–Menten + offset | 10 | 2.8 pp | 2 |
| [PostTrainBench v1.1 → VulcanBench Frontier v4](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1-1/current/to/vulcanbench/current.md) | Hill | 5 | 2.8 pp | 4 |
| [VideoMMMU → MLVU (M-Avg)](https://benchgap.net/calibration/videommmu/current/to/mlvuavg/current.md) | linear | 10 | 2.8 pp | 0 |
| [Gert Labs → AA ITBench](https://benchgap.net/calibration/gertlabs/current/to/aa-itbench/current.md) | Hill | 8 | 2.8 pp | 43 |
| [AA-IFBench → IFEval](https://benchgap.net/calibration/aa-if-bench/current/to/ifeval/current.md) | Hill | 31 | 2.8 pp | 109 |
| [CursorBench 3.1 → Vals LiveCodeBench](https://benchgap.net/calibration/cursor-bench31/current/to/valslivecodebench/current.md) | linear | 6 | 2.8 pp | 1 |
| [MLVU (M-Avg) → Video-MME (with subtitle)](https://benchgap.net/calibration/mlvuavg/current/to/videommewithsub/current.md) | linear | 10 | 2.8 pp | 0 |
| [AA Coding Index → OpenHarmony Bench v1.0](https://benchgap.net/calibration/aa-coding-index/current/to/openharmonybench%23openharmonybench-v1-0/current.md) | linear | 10 | 2.8 pp | 8 |
| [CountBench → AA-MMMU-Pro](https://benchgap.net/calibration/countbench/current/to/aa-mmmu-pro/current.md) | Michaelis–Menten | 5 | 2.8 pp | 1 |
| [PostTrainBench v1 → Vals SWE-bench](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1/current/to/valsswebench/current.md) | Michaelis–Menten | 10 | 2.8 pp | 3 |
| [AA-AnalystAgent → Toolathlon-Verified](https://benchgap.net/calibration/aa-analyst-agent/current/to/toolathlonverified/current.md) | linear | 6 | 2.8 pp | 3 |
| [MLVU (M-Avg) → RealWorldQA](https://benchgap.net/calibration/mlvuavg/current/to/realworldqa/current.md) | Michaelis–Menten + offset | 10 | 2.8 pp | 0 |
| [MMLU-Pro → BBH](https://benchgap.net/calibration/mmlu-pro/current/to/bbh/current.md) | Michaelis–Menten + offset | 8 | 2.8 pp | 52 |
| [FrontierCode 1.1 Main → Vals SWE-bench](https://benchgap.net/calibration/frontiercode/current/to/valsswebench/current.md) | Hill | 9 | 2.8 pp | 4 |
| [PostTrainBench v1.1 → FrontierCode 1.1 Main](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1-1/current/to/frontiercode/current.md) | offset logistic | 6 | 2.8 pp | 4 |
| [RefCOCO (avg) → AI2D_TEST](https://benchgap.net/calibration/refcocoavg/current/to/ai2dtest/current.md) | offset logistic | 6 | 2.8 pp | 0 |
| [We-Math → OmniDocBench 1.5](https://benchgap.net/calibration/wemath/current/to/omnidocbench15/current.md) | offset logistic | 5 | 2.8 pp | 0 |
| [PostTrainBench → VulcanBench Frontier v4](https://benchgap.net/calibration/post-train-bench12/current/to/vulcanbench/current.md) | Michaelis–Menten | 5 | 2.8 pp | 3 |
| [GPQA Diamond → MMLU-Redux](https://benchgap.net/calibration/gpqa-diamond/current/to/mmluredux/current.md) | Michaelis–Menten | 9 | 2.8 pp | 12 |
| [Mystery Game Puzzles → BlenderWorld (text-only)](https://benchgap.net/calibration/mysterygamepuzzles/current/to/blenderworld-text-only/current.md) | Hill | 5 | 2.8 pp | 71 |
| [SpireClimb → ExploitBench Internal Port](https://benchgap.net/calibration/spireclimb/current/to/exploitbenchinternalport/current.md) | inverse Michaelis–Menten | 6 | 2.8 pp | 0 |
| [AA-HLE → BioBasicsBench](https://benchgap.net/calibration/aa-hle/current/to/biobasicsbench/current.md) | offset logistic | 7 | 2.8 pp | 80 |
| [Harvey LAB-AA v1.1 → GDP.pdf](https://benchgap.net/calibration/aa-harvey-lab-v11/current/to/aa-gdp-pdf/current.md) | linear | 6 | 2.8 pp | 4 |
| [MMMU-Pro → RealWorldQA](https://benchgap.net/calibration/mmmu-pro/current/to/realworldqa/current.md) | Michaelis–Menten + offset | 12 | 2.8 pp | 23 |
| [Terminal-Bench 4.1 → Terminal-Bench Hard](https://benchgap.net/calibration/terminal-bench41/current/to/terminal-bench-hard/current.md) | linear | 5 | 2.8 pp | 45 |
| [HMMT Feb 2026 → HMMT Nov 2025](https://benchgap.net/calibration/hmmtfeb2026/current/to/hmmtnov2025/current.md) | Hill | 10 | 2.8 pp | 10 |
| [DynaMath → CountBench](https://benchgap.net/calibration/dynamath/current/to/countbench/current.md) | Michaelis–Menten | 7 | 2.8 pp | 0 |
| [Video-MME (with subtitle) → RealWorldQA](https://benchgap.net/calibration/videommewithsub/current/to/realworldqa/current.md) | offset logistic | 10 | 2.8 pp | 0 |
| [HealthBench Professional (raw, Artificial Analysis) → CritPt](https://benchgap.net/calibration/aa-health-bench-professional-raw/current/to/critpt/current.md) | Hill | 6 | 2.8 pp | 0 |
| [OmniDocBench 1.5 → RealWorldQA](https://benchgap.net/calibration/omnidocbench15/current/to/realworldqa/current.md) | Michaelis–Menten + offset | 9 | 2.9 pp | 0 |
| [CharXiv → Video-MME (with subtitle)](https://benchgap.net/calibration/charxiv/current/to/videommewithsub/current.md) | Michaelis–Menten | 10 | 2.9 pp | 25 |
| [When2Call (Decision Index) → Artificial Analysis Decision Index](https://benchgap.net/calibration/aa-decision-when2-call/current/to/aa-decision-index-v2/current.md) | Hill | 14 | 2.9 pp | 2 |
| [CharXiv w/o tools → ERQA](https://benchgap.net/calibration/charxivnotools/current/to/erqa/current.md) | inverse Michaelis–Menten | 5 | 2.9 pp | 13 |
| [PinchBench → BFCL v4](https://benchgap.net/calibration/pinchbench/current/to/bfcl-v4/current.md) | Michaelis–Menten | 5 | 2.9 pp | 2 |
| [Video-MME (with subtitle) → ERQA](https://benchgap.net/calibration/videommewithsub/current/to/erqa/current.md) | offset logistic | 8 | 2.9 pp | 3 |
| [BrowseComp → DRACO](https://benchgap.net/calibration/browsecomp/current/to/draco/current.md) | linear | 5 | 2.9 pp | 25 |
| [AA-Omniscience Accuracy → Vals MMLU-Pro](https://benchgap.net/calibration/omniscienceaccuracy/current/to/vals-mmlu-pro/current.md) | Hill | 68 | 2.9 pp | 44 |
| [OCRBench V2 → RealWorldQA](https://benchgap.net/calibration/ocrbenchv2/current/to/realworldqa/current.md) | linear | 6 | 2.9 pp | 0 |
| [GDPval-AA → Hermes Bench](https://benchgap.net/calibration/gdpvalaanormalized/current/to/hermesbench/current.md) | Michaelis–Menten | 13 | 2.9 pp | 86 |
| [OSWorld 2.0 → APEX-Agents-AA](https://benchgap.net/calibration/osworld2/current/to/apexagentsaa/current.md) | Michaelis–Menten + offset | 5 | 2.9 pp | 19 |
| [Vals LiveCodeBench → SciCode](https://benchgap.net/calibration/valslivecodebench/current/to/scicode/current.md) | Hill | 7 | 2.9 pp | 53 |
| [Vibe Code Bench → SWE-bench Pro](https://benchgap.net/calibration/vibecodebench/current/to/swe-bench-pro/current.md) | linear | 14 | 2.9 pp | 0 |
| [CountBench → DynaMath](https://benchgap.net/calibration/countbench/current/to/dynamath/current.md) | Michaelis–Menten | 7 | 2.9 pp | 0 |
| [AA-Omniscience Accuracy → GPQA Diamond](https://benchgap.net/calibration/omniscienceaccuracy/current/to/gpqa-diamond/current.md) | Michaelis–Menten + offset | 46 | 2.9 pp | 0 |
| [BabyVision w/ Python → CharXiv w/o tools](https://benchgap.net/calibration/babyvisionpython/current/to/charxivnotools/current.md) | Michaelis–Menten + offset | 6 | 2.9 pp | 0 |
| [DeepSearchQA → ResearchClawBench](https://benchgap.net/calibration/deepsearchqa/current/to/researchclawbench/current.md) | Michaelis–Menten | 6 | 2.9 pp | 1 |
| [AA-MMMU-Pro → RealWorldQA](https://benchgap.net/calibration/aa-mmmu-pro/current/to/realworldqa/current.md) | Michaelis–Menten + offset | 9 | 2.9 pp | 38 |
| [HLE-MultiChoice (Decision Index) → Artificial Analysis Decision Index](https://benchgap.net/calibration/aa-decision-hle-multi-choice/current/to/aa-decision-index-v2/current.md) | offset logistic | 14 | 2.9 pp | 0 |
| [MMVU → MMMU](https://benchgap.net/calibration/mmvu/current/to/mmmu/current.md) | Hill | 5 | 2.9 pp | 1 |
| [Vals GPQA Diamond → Vals MMLU-Pro](https://benchgap.net/calibration/vals-gpqa-diamond/current/to/vals-mmlu-pro/current.md) | linear | 85 | 2.9 pp | 0 |
| [BlenderWorld (text-only) → Cookie-Bench](https://benchgap.net/calibration/blenderworld-text-only/current/to/cookie-bench/current.md) | Michaelis–Menten + offset | 7 | 2.9 pp | 0 |
| [V* → RealWorldQA](https://benchgap.net/calibration/vstar/current/to/realworldqa/current.md) | Michaelis–Menten | 6 | 2.9 pp | 0 |
| [MMMLU → CritPt](https://benchgap.net/calibration/mmmlu/current/to/critpt/current.md) | linear | 6 | 2.9 pp | 3 |
| [ZeroBench → CharXiv w/o tools](https://benchgap.net/calibration/zerobench/current/to/charxivnotools/current.md) | Michaelis–Menten | 5 | 2.9 pp | 0 |
| [AIME 2025 → HMMT Feb 2025](https://benchgap.net/calibration/aime2025/current/to/hmmtfeb2025/current.md) | offset logistic | 8 | 2.9 pp | 25 |
| [LifeVal → BrowseComp](https://benchgap.net/calibration/lifeval/current/to/browsecomp/current.md) | Michaelis–Menten | 5 | 2.9 pp | 2 |
| [Chartography (tools) → OfficeQA Pro](https://benchgap.net/calibration/chartographywithtools/current/to/officeqapro/current.md) | Michaelis–Menten | 7 | 2.9 pp | 3 |
| [HLE w/o tools → Vals GPQA Diamond](https://benchgap.net/calibration/hlenotools/current/to/vals-gpqa-diamond/current.md) | Michaelis–Menten | 21 | 2.9 pp | 7 |
| [CharXiv w/o tools → AA-MMMU-Pro](https://benchgap.net/calibration/charxivnotools/current/to/aa-mmmu-pro/current.md) | linear | 11 | 2.9 pp | 1 |
| [AA-MMMU-Pro → CountBench](https://benchgap.net/calibration/aa-mmmu-pro/current/to/countbench/current.md) | Michaelis–Menten | 5 | 2.9 pp | 59 |
| [HLE-Verified → ARC-AGI-2](https://benchgap.net/calibration/hleverified/current/to/arc-agi-2/current.md) | Hill | 5 | 2.9 pp | 0 |
| [ApprenticeBench → BrowseComp](https://benchgap.net/calibration/apprenticebench/current/to/browsecomp/current.md) | Hill | 11 | 2.9 pp | 2 |
| [Bug Hunt Bench → AA Coding Index](https://benchgap.net/calibration/bug-hunt-bench/current/to/aa-coding-index/current.md) | Hill | 10 | 2.9 pp | 0 |
| [AI2D_TEST → CharXiv](https://benchgap.net/calibration/ai2dtest/current/to/charxiv/current.md) | inverse Michaelis–Menten | 7 | 3.0 pp | 1 |
| [ApprenticeBench → Gert Labs](https://benchgap.net/calibration/apprenticebench/current/to/gertlabs/current.md) | linear | 5 | 3.0 pp | 11 |
| [SWE-bench Pro → FrontierCode 1.1 Extended](https://benchgap.net/calibration/swe-bench-pro/current/to/frontiercode11extended/current.md) | Michaelis–Menten | 6 | 3.0 pp | 64 |
| [PostTrainBench → FrontierCode 1.1 Main](https://benchgap.net/calibration/post-train-bench12/current/to/frontiercode/current.md) | Michaelis–Menten | 6 | 3.0 pp | 10 |
| [SuperGPQA → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/supergpqa/current/to/artificialanalysis/current.md) | offset logistic | 12 | 3.0 pp | 9 |
| [MMMU-Pro → MathVision](https://benchgap.net/calibration/mmmu-pro/current/to/mathvision/current.md) | linear | 20 | 3.0 pp | 35 |
| [MMMU-Pro → CountBench](https://benchgap.net/calibration/mmmu-pro/current/to/countbench/current.md) | Michaelis–Menten | 8 | 3.0 pp | 11 |
| [Decision Index Arts → When2Call (Decision Index)](https://benchgap.net/calibration/jevdecisionindex03arts/current/to/aa-decision-when2-call/current.md) | Hill | 11 | 3.0 pp | 72 |
| [Vibe Code Bench → React Native Evals](https://benchgap.net/calibration/vibecodebench/current/to/reactnativeevals/current.md) | Michaelis–Menten | 7 | 3.0 pp | 25 |
| [ZeroBench w/ Python → BabyVision w/ Python](https://benchgap.net/calibration/zerobenchpython/current/to/babyvisionpython/current.md) | linear | 5 | 3.0 pp | 2 |
| [CursorBench 3.2 → CursorBench 4.0](https://benchgap.net/calibration/cursor-bench32/current/to/cursor-bench40/current.md) | linear | 8 | 3.0 pp | 1 |
| [AA-MMMU-Pro → DynaMath](https://benchgap.net/calibration/aa-mmmu-pro/current/to/dynamath/current.md) | offset logistic | 5 | 3.0 pp | 51 |
| [FrontierSWE v2 → AA Coding Index](https://benchgap.net/calibration/frontierswev2/current/to/aa-coding-index/current.md) | Hill | 13 | 3.0 pp | 0 |
| [MathVision → MMMU-Pro](https://benchgap.net/calibration/mathvision/current/to/mmmu-pro/current.md) | Michaelis–Menten + offset | 20 | 3.0 pp | 1 |
| [CursorBench 4.0 → FrontierCode 1.1 Extended](https://benchgap.net/calibration/cursor-bench40/current/to/frontiercode11extended/current.md) | Hill | 7 | 3.0 pp | 8 |
| [PostTrainBench v1.1 → AA Coding Index](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1-1/current/to/aa-coding-index/current.md) | linear | 12 | 3.0 pp | 0 |
| [MMMU-Pro → Video-MME (with subtitle)](https://benchgap.net/calibration/mmmu-pro/current/to/videommewithsub/current.md) | Michaelis–Menten | 12 | 3.0 pp | 25 |
| [MMMU → AA-MMMU-Pro](https://benchgap.net/calibration/mmmu/current/to/aa-mmmu-pro/current.md) | offset logistic | 15 | 3.0 pp | 8 |
| [DynaMath → AA-MMMU-Pro](https://benchgap.net/calibration/dynamath/current/to/aa-mmmu-pro/current.md) | inverse Michaelis–Menten | 5 | 3.0 pp | 0 |
| [ExtractBench Pro → AA-MMMU-Pro](https://benchgap.net/calibration/extractbench-pro/current/to/aa-mmmu-pro/current.md) | Michaelis–Menten + offset | 5 | 3.0 pp | 2 |
| [AA-GPQA Diamond → Vals MMLU-Pro](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/vals-mmlu-pro/current.md) | Michaelis–Menten | 74 | 3.0 pp | 25 |
| [PerceptionBench → MMMU-Pro](https://benchgap.net/calibration/perceptionbench/current/to/mmmu-pro/current.md) | linear | 6 | 3.0 pp | 0 |
| [SEC-Bench Pro → FACTS Multimodal](https://benchgap.net/calibration/secbenchpro/current/to/facts-multimodal/current.md) | Hill | 8 | 3.0 pp | 2 |
| [ExploitGym → BrowseComp](https://benchgap.net/calibration/exploitgym/current/to/browsecomp/current.md) | linear | 7 | 3.0 pp | 12 |
| [Artificial Analysis Intelligence Index → MMMLU](https://benchgap.net/calibration/artificialanalysis/current/to/mmmlu/current.md) | linear | 7 | 3.0 pp | 3 |
| [OSWorld 2.0 → OSWorld-Verified](https://benchgap.net/calibration/osworld2/current/to/osworld-verified/current.md) | linear | 9 | 3.0 pp | 1 |
| [SWE-Marathon → AA-SciCode](https://benchgap.net/calibration/swemarathon/current/to/aa-sci-code/current.md) | offset logistic | 7 | 3.0 pp | 2 |
| [CursorBench 4.0 → Vals LiveCodeBench](https://benchgap.net/calibration/cursor-bench40/current/to/valslivecodebench/current.md) | linear | 8 | 3.0 pp | 9 |
| [SciCode → AA-SciCode](https://benchgap.net/calibration/scicode/current/to/aa-sci-code/current.md) | linear | 20 | 3.0 pp | 15 |
| [CursorBench 3.2 → AA Coding Index](https://benchgap.net/calibration/cursor-bench32/current/to/aa-coding-index/current.md) | linear | 17 | 3.0 pp | 0 |
| [GDPval-AA → Toolathlon-Verified](https://benchgap.net/calibration/gdpvalaanormalized/current/to/toolathlonverified/current.md) | Michaelis–Menten | 14 | 3.1 pp | 41 |
| [Hermes Bench → HLE w/ tools](https://benchgap.net/calibration/hermesbench/current/to/hlewithtools/current.md) | inverse Michaelis–Menten | 8 | 3.1 pp | 7 |
| [BioBasicsBench → ARC-AGI-3](https://benchgap.net/calibration/biobasicsbench/current/to/arcagi3/current.md) | Hill | 6 | 3.1 pp | 1 |
| [Toolathlon-Verified → DRACO](https://benchgap.net/calibration/toolathlonverified/current/to/draco/current.md) | Michaelis–Menten | 7 | 3.1 pp | 0 |
| [Vibe Code Bench → PostTrainBench v1](https://benchgap.net/calibration/vibecodebench/current/to/posttrainbench%23posttrainbench-v1/current.md) | Hill | 10 | 3.1 pp | 29 |
| [Gert Labs → ApprenticeBench](https://benchgap.net/calibration/gertlabs/current/to/apprenticebench/current.md) | linear | 5 | 3.1 pp | 15 |
| [CharXiv → MMMU-Pro](https://benchgap.net/calibration/charxiv/current/to/mmmu-pro/current.md) | Michaelis–Menten | 30 | 3.1 pp | 7 |
| [HLE w/ tools → τ²-bench results](https://benchgap.net/calibration/hlewithtools/current/to/tau2-bench/current.md) | Hill | 8 | 3.1 pp | 29 |
| [AA Tau3 Banking → CyberGym](https://benchgap.net/calibration/aa-tau3-banking/current/to/cybergym/current.md) | Michaelis–Menten | 7 | 3.1 pp | 13 |
| [AA-MMMU-Pro → ERQA](https://benchgap.net/calibration/aa-mmmu-pro/current/to/erqa/current.md) | offset logistic | 11 | 3.1 pp | 65 |
| [AA-MMMU-Pro → MathVision](https://benchgap.net/calibration/aa-mmmu-pro/current/to/mathvision/current.md) | Hill | 15 | 3.1 pp | 49 |
| [ExploitGym → MKA](https://benchgap.net/calibration/exploitgym/current/to/mka/current.md) | Michaelis–Menten | 8 | 3.1 pp | 14 |
| [FrontierMath v2 (Tiers 1-3) → HMMT Feb 2026](https://benchgap.net/calibration/frontiermathv2tiers13/current/to/hmmtfeb2026/current.md) | linear | 11 | 3.1 pp | 27 |
| [ARC-AGI-2 → LABBench2](https://benchgap.net/calibration/arc-agi-2/current/to/labbench2/current.md) | Michaelis–Menten | 6 | 3.1 pp | 1 |
| [VulcanBench Frontier v4 → CursorBench 3.2](https://benchgap.net/calibration/vulcanbench/current/to/cursor-bench32/current.md) | inverse Michaelis–Menten | 5 | 3.1 pp | 6 |
| [RealWorldQA → RefCOCO (avg)](https://benchgap.net/calibration/realworldqa/current/to/refcocoavg/current.md) | Michaelis–Menten | 8 | 3.1 pp | 9 |
| [AIME26 → MATH-500](https://benchgap.net/calibration/aime2026/current/to/math-500/current.md) | linear | 5 | 3.1 pp | 26 |
| [Artificial Analysis Decision Index v1.0 → Artificial Analysis Decision Index](https://benchgap.net/calibration/aa-decision-index/current/to/aa-decision-index-v2/current.md) | linear | 14 | 3.1 pp | 0 |
| [AA MMLU-Pro → Vals GPQA Diamond](https://benchgap.net/calibration/aa-mmlu-pro/current/to/vals-gpqa-diamond/current.md) | inverse Michaelis–Menten | 10 | 3.1 pp | 3 |
| [MazeBench → Hermes Index](https://benchgap.net/calibration/mazebench/current/to/hermesindex/current.md) | offset logistic | 13 | 3.1 pp | 30 |
| [AA-SciCode → SciCode](https://benchgap.net/calibration/aa-sci-code/current/to/scicode/current.md) | linear | 20 | 3.1 pp | 66 |
| [GDPval-AA → AA Agentic Index](https://benchgap.net/calibration/gdpvalaanormalized/current/to/aa-agentic-index/current.md) | inverse Michaelis–Menten | 69 | 3.1 pp | 61 |
| [MKA → GDP.pdf](https://benchgap.net/calibration/mka/current/to/aa-gdp-pdf/current.md) | Hill | 15 | 3.1 pp | 5 |
| [NL2Repo → AA-SciCode](https://benchgap.net/calibration/nl2repo/current/to/aa-sci-code/current.md) | Michaelis–Menten + offset | 16 | 3.1 pp | 16 |
| [BabyVision → CharXiv](https://benchgap.net/calibration/babyvision/current/to/charxiv/current.md) | Michaelis–Menten | 10 | 3.2 pp | 1 |
| [CWE-bench → MKA](https://benchgap.net/calibration/cwe-bench-v2/current/to/mka/current.md) | Michaelis–Menten | 14 | 3.2 pp | 3 |
| [MMMU → CountBench](https://benchgap.net/calibration/mmmu/current/to/countbench/current.md) | Hill | 10 | 3.2 pp | 10 |
| [AA Coding Index → React Native Evals](https://benchgap.net/calibration/aa-coding-index/current/to/reactnativeevals/current.md) | linear | 9 | 3.2 pp | 54 |
| [Artificial Analysis Decision Index v1.0 → When2Call (Decision Index)](https://benchgap.net/calibration/aa-decision-index/current/to/aa-decision-when2-call/current.md) | Hill | 19 | 3.2 pp | 0 |
| [StealthMark → DAYJOB](https://benchgap.net/calibration/stealthmark/current/to/dayjob/current.md) | linear | 12 | 3.2 pp | 18 |
| [VulcanBench Frontier v4 → AA Coding Index](https://benchgap.net/calibration/vulcanbench/current/to/aa-coding-index/current.md) | Michaelis–Menten + offset | 6 | 3.2 pp | 1 |
| [AIME26 → HMMT Nov 2025](https://benchgap.net/calibration/aime2026/current/to/hmmtnov2025/current.md) | Hill | 10 | 3.2 pp | 7 |
| [Mystery Game Puzzles → DAYJOB](https://benchgap.net/calibration/mysterygamepuzzles/current/to/dayjob/current.md) | offset logistic | 19 | 3.2 pp | 26 |
| [PostTrainBench v1 → SWE-bench Verified](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1/current/to/swe-bench-verified/current.md) | linear | 6 | 3.2 pp | 5 |
| [MMLU-ProX → SWE Multilingual](https://benchgap.net/calibration/mmluprox/current/to/swe-bench-multilingual/current.md) | inverse Michaelis–Menten | 8 | 3.2 pp | 4 |
| [SimpleVQA → MMMU](https://benchgap.net/calibration/simplevqa/current/to/mmmu/current.md) | Hill | 9 | 3.2 pp | 0 |
| [BabyVision w/ Python → AA-MMMU-Pro](https://benchgap.net/calibration/babyvisionpython/current/to/aa-mmmu-pro/current.md) | Michaelis–Menten + offset | 6 | 3.2 pp | 0 |
| [AA-GPQA Diamond → Vals GPQA Diamond](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/vals-gpqa-diamond/current.md) | linear | 74 | 3.2 pp | 100 |
| [MMMU → Video-MME (with subtitle)](https://benchgap.net/calibration/mmmu/current/to/videommewithsub/current.md) | Michaelis–Menten | 8 | 3.2 pp | 12 |
| [Decision Index Arts → Artificial Analysis Decision Index v1.0](https://benchgap.net/calibration/jevdecisionindex03arts/current/to/aa-decision-index/current.md) | Michaelis–Menten | 11 | 3.2 pp | 72 |
| [SWE-Marathon → Bug Hunt Bench](https://benchgap.net/calibration/swemarathon/current/to/bug-hunt-bench/current.md) | linear | 5 | 3.2 pp | 4 |
| [ResearchClawBench → Claw-Eval](https://benchgap.net/calibration/researchclawbench/current/to/claw-eval/current.md) | Hill | 12 | 3.2 pp | 2 |
| [HLE w/o tools → GPQA](https://benchgap.net/calibration/hlenotools/current/to/gpqa/current.md) | Hill | 30 | 3.2 pp | 3 |
| [Vals SWE-bench → OpenHarmony Bench v1.0](https://benchgap.net/calibration/valsswebench/current/to/openharmonybench%23openharmonybench-v1-0/current.md) | linear | 11 | 3.2 pp | 20 |
| [Terminal-Bench 4.1 → Terminal-Bench 3.0](https://benchgap.net/calibration/terminal-bench41/current/to/terminal-bench%23terminal-bench-3-0/current.md) | linear | 10 | 3.2 pp | 35 |
| [MMLU → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/mmlu/current/to/artificialanalysis/current.md) | linear | 14 | 3.2 pp | 17 |
| [RealWorldQA → CharXiv](https://benchgap.net/calibration/realworldqa/current/to/charxiv/current.md) | inverse Michaelis–Menten | 12 | 3.3 pp | 3 |
| [CyberGym → AA Tau3 Banking](https://benchgap.net/calibration/cybergym/current/to/aa-tau3-banking/current.md) | offset logistic | 7 | 3.3 pp | 10 |
| [VulcanBench v3 → CursorBench 3.2](https://benchgap.net/calibration/vulcanbench%23vulcanbench-v3/current/to/cursor-bench32/current.md) | Michaelis–Menten | 8 | 3.3 pp | 5 |
| [MathVision → MMVU](https://benchgap.net/calibration/mathvision/current/to/mmvu/current.md) | linear | 7 | 3.3 pp | 5 |
| [ExploitBench Internal Port → FACTS Multimodal](https://benchgap.net/calibration/exploitbenchinternalport/current/to/facts-multimodal/current.md) | linear | 6 | 3.3 pp | 0 |
| [SWE-bench Pro → CursorBench 4.0](https://benchgap.net/calibration/swe-bench-pro/current/to/cursor-bench40/current.md) | linear | 9 | 3.3 pp | 53 |
| [CharXiv → MathVision](https://benchgap.net/calibration/charxiv/current/to/mathvision/current.md) | Michaelis–Menten | 22 | 3.3 pp | 6 |
| [Decision Index Knowledge → Decision Index Full Score](https://benchgap.net/calibration/jevdecisionindex03knowledge/current/to/decisionindex/current.md) | Hill | 110 | 3.3 pp | 0 |
| [VideoMMMU → RealWorldQA](https://benchgap.net/calibration/videommmu/current/to/realworldqa/current.md) | Hill | 11 | 3.3 pp | 0 |
| [AutomationBench → AA Tau3 Banking](https://benchgap.net/calibration/automationbench/current/to/aa-tau3-banking/current.md) | Michaelis–Menten | 7 | 3.3 pp | 16 |
| [AA-Omniscience Accuracy → MMMLU](https://benchgap.net/calibration/omniscienceaccuracy/current/to/mmmlu/current.md) | offset logistic | 5 | 3.3 pp | 0 |
| [PostTrainBench v1.1 → CursorBench 4.0](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1-1/current/to/cursor-bench40/current.md) | Michaelis–Menten | 5 | 3.3 pp | 1 |
| [AA Agentic Index → GDPval-AA](https://benchgap.net/calibration/aa-agentic-index/current/to/gdpvalaanormalized/current.md) | Michaelis–Menten | 69 | 3.3 pp | 0 |
| [VulcanBench Frontier v4 → PostTrainBench](https://benchgap.net/calibration/vulcanbench/current/to/post-train-bench12/current.md) | Hill | 5 | 3.3 pp | 8 |
| [MMMU-Pro → OCRBench V2](https://benchgap.net/calibration/mmmu-pro/current/to/ocrbenchv2/current.md) | Michaelis–Menten | 8 | 3.3 pp | 41 |
| [FrontierCode 2.0 → CursorBench 4.0](https://benchgap.net/calibration/frontiercode-2/current/to/cursor-bench40/current.md) | Hill | 13 | 3.3 pp | 13 |
| [Toolathlon-Verified → AA Agentic Index](https://benchgap.net/calibration/toolathlonverified/current/to/aa-agentic-index/current.md) | linear | 7 | 3.3 pp | 9 |
| [CyberGym → Agents' Last Exam](https://benchgap.net/calibration/cybergym/current/to/agentslastexam/current.md) | Michaelis–Menten + offset | 9 | 3.3 pp | 15 |
| [AA-HLE → MMLU-Redux](https://benchgap.net/calibration/aa-hle/current/to/mmluredux/current.md) | Michaelis–Menten | 9 | 3.3 pp | 19 |
| [HealthBench (raw) → HealthBench Professional](https://benchgap.net/calibration/healthbench/current/to/healthbenchprofessional/current.md) | Michaelis–Menten | 8 | 3.4 pp | 0 |
| [ARC-AGI-3 Provider Adapter → ARC-AGI-1](https://benchgap.net/calibration/arc-agi-3-provider-adapter/current/to/arcagi1/current.md) | linear | 12 | 3.4 pp | 1 |
| [MMLU-Pro → MMMLU](https://benchgap.net/calibration/mmlu-pro/current/to/mmmlu/current.md) | Michaelis–Menten | 5 | 3.4 pp | 1 |
| [OpenHarmony Bench v1.0 → AA Coding Index](https://benchgap.net/calibration/openharmonybench%23openharmonybench-v1-0/current/to/aa-coding-index/current.md) | offset logistic | 10 | 3.4 pp | 0 |
| [We-Math → MLVU (M-Avg)](https://benchgap.net/calibration/wemath/current/to/mlvuavg/current.md) | offset logistic | 6 | 3.4 pp | 0 |
| [LiveCodeBench v6 → NL2Repo](https://benchgap.net/calibration/livecodebench-v6/current/to/nl2repo/current.md) | Hill | 12 | 3.4 pp | 23 |
| [MCP-Tasks → Claw-Eval](https://benchgap.net/calibration/mcptasks/current/to/claw-eval/current.md) | offset logistic | 5 | 3.4 pp | 0 |
| [CWE-Bench-AA → Hermes Bench](https://benchgap.net/calibration/aa-cwe-bench/current/to/hermesbench/current.md) | Hill | 10 | 3.4 pp | 0 |
| [CritPt → GPQA Diamond](https://benchgap.net/calibration/critpt/current/to/gpqa-diamond/current.md) | Michaelis–Menten + offset | 52 | 3.4 pp | 1 |
| [HMMT Nov 2025 → HMMT Feb 2026](https://benchgap.net/calibration/hmmtnov2025/current/to/hmmtfeb2026/current.md) | Hill | 10 | 3.4 pp | 1 |
| [Vibe Code Bench → AA LiveCodeBench](https://benchgap.net/calibration/vibecodebench/current/to/aa-live-code-bench/current.md) | linear | 7 | 3.4 pp | 8 |
| [LiveCodeBench v6 → Vals LiveCodeBench](https://benchgap.net/calibration/livecodebench-v6/current/to/valslivecodebench/current.md) | offset logistic | 7 | 3.4 pp | 19 |
| [CWE-bench → Hermes Index](https://benchgap.net/calibration/cwe-bench-v2/current/to/hermesindex/current.md) | Hill | 7 | 3.4 pp | 1 |
| [Vals SWE-bench → CursorBench 4.0](https://benchgap.net/calibration/valsswebench/current/to/cursor-bench40/current.md) | offset logistic | 9 | 3.4 pp | 18 |
| [AA-HLE → HealthBench (length-adjusted)](https://benchgap.net/calibration/aa-hle/current/to/healthbenchlengthadjusted/current.md) | linear | 8 | 3.4 pp | 80 |
| [VulcanBench v3 → FrontierCode 1.1 Extended](https://benchgap.net/calibration/vulcanbench%23vulcanbench-v3/current/to/frontiercode11extended/current.md) | offset logistic | 5 | 3.4 pp | 1 |
| [AA-SciCode → CursorBench 3.1](https://benchgap.net/calibration/aa-sci-code/current/to/cursor-bench31/current.md) | linear | 5 | 3.4 pp | 73 |
| [MRCRv2 → AA-LCR](https://benchgap.net/calibration/mrcrv2/current/to/lcr/current.md) | Hill | 9 | 3.4 pp | 5 |
| [OpenHarmony Bench v1.0 → AA-SciCode](https://benchgap.net/calibration/openharmonybench%23openharmonybench-v1-0/current/to/aa-sci-code/current.md) | Michaelis–Menten | 11 | 3.4 pp | 0 |
| [AA-HLE → CritPt](https://benchgap.net/calibration/aa-hle/current/to/critpt/current.md) | offset logistic | 200 | 3.4 pp | 9 |
| [GPQA Diamond → MMLU-Pro (Arcee)](https://benchgap.net/calibration/gpqa-diamond/current/to/mmlu-pro-arcee/current.md) | Michaelis–Menten | 6 | 3.4 pp | 80 |
| [ARC-AGI-3 → ARC-AGI-1](https://benchgap.net/calibration/arcagi3/current/to/arcagi1/current.md) | offset logistic | 21 | 3.4 pp | 0 |
| [AA Coding Index → FrontierCode 1.1 Extended](https://benchgap.net/calibration/aa-coding-index/current/to/frontiercode11extended/current.md) | Michaelis–Menten + offset | 6 | 3.4 pp | 3 |
| [AA Coding Index → AA-SciCode](https://benchgap.net/calibration/aa-coding-index/current/to/aa-sci-code/current.md) | linear | 74 | 3.4 pp | 20 |
| [MLVU (M-Avg) → RefCOCO (avg)](https://benchgap.net/calibration/mlvuavg/current/to/refcocoavg/current.md) | offset logistic | 6 | 3.4 pp | 0 |
| [OSWorld 2.1 Offline → MKA](https://benchgap.net/calibration/osworld-2-1-offline/current/to/mka/current.md) | offset logistic | 9 | 3.4 pp | 2 |
| [AIME 2024 → AIME 2025](https://benchgap.net/calibration/aime2024/current/to/aime2025/current.md) | inverse Michaelis–Menten | 17 | 3.5 pp | 11 |
| [NL2Repo → SWE-bench Pro](https://benchgap.net/calibration/nl2repo/current/to/swe-bench-pro/current.md) | Hill | 28 | 3.5 pp | 8 |
| [ApprenticeBench → APEX-Agents-AA](https://benchgap.net/calibration/apprenticebench/current/to/apexagentsaa/current.md) | linear | 5 | 3.5 pp | 3 |
| [MLVU (M-Avg) → MathVision](https://benchgap.net/calibration/mlvuavg/current/to/mathvision/current.md) | Hill | 8 | 3.5 pp | 0 |
| [MKA → Hermes Bench](https://benchgap.net/calibration/mka/current/to/hermesbench/current.md) | linear | 10 | 3.5 pp | 0 |
| [ARC-AGI-3 → BioBasicsBench](https://benchgap.net/calibration/arcagi3/current/to/biobasicsbench/current.md) | linear | 6 | 3.5 pp | 3 |
| [WildGuardTest (Decision Index) → Artificial Analysis Decision Index](https://benchgap.net/calibration/aa-decision-wild-guard-test/current/to/aa-decision-index-v2/current.md) | inverse Michaelis–Menten | 14 | 3.5 pp | 0 |
| [BrowseComp → WideResearch](https://benchgap.net/calibration/browsecomp/current/to/wideresearch/current.md) | Michaelis–Menten | 11 | 3.5 pp | 29 |
| [Video-MME (w/o subtitle) → Video-MME (with subtitle)](https://benchgap.net/calibration/videommenosub/current/to/videommewithsub/current.md) | Michaelis–Menten | 7 | 3.5 pp | 0 |
| [GDP.pdf → Harvey LAB-AA v1.1](https://benchgap.net/calibration/aa-gdp-pdf/current/to/aa-harvey-lab-v11/current.md) | linear | 6 | 3.5 pp | 15 |
| [FrontierCode 2.0 → AA-SciCode](https://benchgap.net/calibration/frontiercode-2/current/to/aa-sci-code/current.md) | Hill | 21 | 3.5 pp | 7 |
| [DeepSWE → FrontierCode 1.1 Main](https://benchgap.net/calibration/deepswe/current/to/frontiercode/current.md) | Michaelis–Menten + offset | 7 | 3.5 pp | 33 |
| [BioBasicsBench → AA-HLE](https://benchgap.net/calibration/biobasicsbench/current/to/aa-hle/current.md) | Michaelis–Menten | 7 | 3.5 pp | 0 |
| [MMMU → RealWorldQA](https://benchgap.net/calibration/mmmu/current/to/realworldqa/current.md) | Michaelis–Menten | 12 | 3.5 pp | 10 |
| [Artificial Analysis Decision Index → Decision Index Knowledge](https://benchgap.net/calibration/aa-decision-index-v2/current/to/jevdecisionindex03knowledge/current.md) | inverse Michaelis–Menten | 11 | 3.5 pp | 11 |
| [Artificial Analysis Decision Index v1.0 → Decision Index Arts](https://benchgap.net/calibration/aa-decision-index/current/to/jevdecisionindex03arts/current.md) | Hill | 11 | 3.5 pp | 0 |
| [DAYJOB → StealthMark](https://benchgap.net/calibration/dayjob/current/to/stealthmark/current.md) | Michaelis–Menten + offset | 12 | 3.5 pp | 5 |
| [HealthBench Professional → Vals MMLU-Pro](https://benchgap.net/calibration/healthbenchprofessional/current/to/vals-mmlu-pro/current.md) | Hill | 6 | 3.5 pp | 1 |
| [PostTrainBench v1.1 → AA-SciCode](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1-1/current/to/aa-sci-code/current.md) | Hill | 13 | 3.5 pp | 0 |
| [FrontierMath v2 (Tier 4) → FrontierMath v2 (Tiers 1-3)](https://benchgap.net/calibration/frontiermathv2tier4/current/to/frontiermathv2tiers13/current.md) | Michaelis–Menten + offset | 58 | 3.5 pp | 0 |
| [Video-MME (with subtitle) → We-Math](https://benchgap.net/calibration/videommewithsub/current/to/wemath/current.md) | offset logistic | 6 | 3.5 pp | 3 |
| [SWE-bench Pro → SWE-bench Verified](https://benchgap.net/calibration/swe-bench-pro/current/to/swe-bench-verified/current.md) | linear | 51 | 3.6 pp | 19 |
| [SWE-bench Pro → CursorBench 3.2](https://benchgap.net/calibration/swe-bench-pro/current/to/cursor-bench32/current.md) | Hill | 12 | 3.6 pp | 61 |
| [PostTrainBench v1 → SWE-bench Pro](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1/current/to/swe-bench-pro/current.md) | offset logistic | 9 | 3.6 pp | 0 |
| [SimpleQA → MMLU-Pro](https://benchgap.net/calibration/simpleqa/current/to/mmlu-pro/current.md) | Michaelis–Menten | 8 | 3.6 pp | 7 |
| [BrowseComp → AA Tau3 Banking](https://benchgap.net/calibration/browsecomp/current/to/aa-tau3-banking/current.md) | inverse Michaelis–Menten | 10 | 3.6 pp | 26 |
| [MMMU → MathVision](https://benchgap.net/calibration/mmmu/current/to/mathvision/current.md) | linear | 10 | 3.6 pp | 10 |
| [CharXiv → MMVU](https://benchgap.net/calibration/charxiv/current/to/mmvu/current.md) | Michaelis–Menten + offset | 6 | 3.6 pp | 14 |
| [MathVision → PerceptionBench](https://benchgap.net/calibration/mathvision/current/to/perceptionbench/current.md) | inverse Michaelis–Menten | 5 | 3.6 pp | 12 |
| [Video-MME (w/o subtitle) → VideoMMMU](https://benchgap.net/calibration/videommenosub/current/to/videommmu/current.md) | inverse Michaelis–Menten | 8 | 3.6 pp | 0 |
| [SWE-bench Verified → SWE-Rebench](https://benchgap.net/calibration/swe-bench-verified/current/to/swe-rebench/current.md) | offset logistic | 8 | 3.6 pp | 84 |
| [IFBench → IFEval](https://benchgap.net/calibration/ifbench/current/to/ifeval/current.md) | Hill | 22 | 3.6 pp | 28 |
| [MathVision → ZeroBench](https://benchgap.net/calibration/mathvision/current/to/zerobench/current.md) | offset logistic | 6 | 3.6 pp | 14 |
| [RealWorldQA → CountBench](https://benchgap.net/calibration/realworldqa/current/to/countbench/current.md) | Michaelis–Menten | 10 | 3.6 pp | 2 |
| [ZeroBench → ScreenSpot Pro](https://benchgap.net/calibration/zerobench/current/to/screenspot-pro/current.md) | Hill | 5 | 3.6 pp | 6 |
| [ZeroBench w/ Python → MMMU-Pro](https://benchgap.net/calibration/zerobenchpython/current/to/mmmu-pro/current.md) | Michaelis–Menten | 6 | 3.6 pp | 0 |
| [AI2D_TEST → Video-MME (with subtitle)](https://benchgap.net/calibration/ai2dtest/current/to/videommewithsub/current.md) | Michaelis–Menten | 6 | 3.6 pp | 0 |
| [Vals SWE-bench → PostTrainBench v1](https://benchgap.net/calibration/valsswebench/current/to/posttrainbench%23posttrainbench-v1/current.md) | linear | 10 | 3.6 pp | 33 |
| [AA MMLU-Pro → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/aa-mmlu-pro/current/to/artificialanalysis/current.md) | linear | 12 | 3.6 pp | 1 |
| [BrowseComp → PinchBench](https://benchgap.net/calibration/browsecomp/current/to/pinchbench/current.md) | Michaelis–Menten | 5 | 3.6 pp | 58 |
| [Hermes Index → Hermes Bench](https://benchgap.net/calibration/hermesindex/current/to/hermesbench/current.md) | Michaelis–Menten | 16 | 3.6 pp | 0 |
| [SWE-bench Verified → SciCode](https://benchgap.net/calibration/swe-bench-verified/current/to/scicode/current.md) | Michaelis–Menten | 20 | 3.6 pp | 65 |
| [Vals GPQA Diamond → AA-GPQA Diamond](https://benchgap.net/calibration/vals-gpqa-diamond/current/to/aa-gpqa-diamond/current.md) | linear | 74 | 3.6 pp | 9 |
| [τ³-bench results → MCP-Tasks](https://benchgap.net/calibration/tau3-bench/current/to/mcptasks/current.md) | Hill | 5 | 3.6 pp | 6 |
| [OSWorld-Verified → OSWorld 2.0](https://benchgap.net/calibration/osworld-verified/current/to/osworld2/current.md) | linear | 9 | 3.6 pp | 0 |
| [ProgramBench → CursorBench 3.2](https://benchgap.net/calibration/programbench/current/to/cursor-bench32/current.md) | Michaelis–Menten | 5 | 3.6 pp | 8 |
| [V* → CharXiv](https://benchgap.net/calibration/vstar/current/to/charxiv/current.md) | offset logistic | 12 | 3.6 pp | 1 |
| [SWE-bench Pro → AA-SciCode](https://benchgap.net/calibration/swe-bench-pro/current/to/aa-sci-code/current.md) | linear | 43 | 3.6 pp | 15 |
| [CursorBench 3.1 → AA Coding Index](https://benchgap.net/calibration/cursor-bench31/current/to/aa-coding-index/current.md) | Hill | 5 | 3.6 pp | 1 |
| [JobBench → APEX-Agents-AA](https://benchgap.net/calibration/jobbench/current/to/apexagentsaa/current.md) | Michaelis–Menten | 5 | 3.6 pp | 25 |
| [ARC-AGI-3 → HealthBench Professional](https://benchgap.net/calibration/arcagi3/current/to/healthbenchprofessional/current.md) | linear | 12 | 3.6 pp | 8 |
| [Hermes Bench → DRACO](https://benchgap.net/calibration/hermesbench/current/to/draco/current.md) | Michaelis–Menten | 5 | 3.7 pp | 0 |
| [CharXiv → CountBench](https://benchgap.net/calibration/charxiv/current/to/countbench/current.md) | Michaelis–Menten | 6 | 3.7 pp | 8 |
| [Artificial Analysis Intelligence Index → HealthBench (length-adjusted)](https://benchgap.net/calibration/artificialanalysis/current/to/healthbenchlengthadjusted/current.md) | offset logistic | 8 | 3.7 pp | 72 |
| [CharXiv w/o tools → BabyVision w/ Python](https://benchgap.net/calibration/charxivnotools/current/to/babyvisionpython/current.md) | Michaelis–Menten | 6 | 3.7 pp | 10 |
| [MathVision → MMMU](https://benchgap.net/calibration/mathvision/current/to/mmmu/current.md) | Hill | 10 | 3.7 pp | 3 |
| [MathVision → BabyVision](https://benchgap.net/calibration/mathvision/current/to/babyvision/current.md) | inverse Michaelis–Menten | 9 | 3.7 pp | 15 |
| [Vals SWE-bench → PostTrainBench](https://benchgap.net/calibration/valsswebench/current/to/post-train-bench12/current.md) | Hill | 12 | 3.7 pp | 40 |
| [OmniDocBench 1.5 → CC-OCR](https://benchgap.net/calibration/omnidocbench15/current/to/ccocr/current.md) | inverse Michaelis–Menten | 6 | 3.7 pp | 4 |
| [AA Agentic Index → AA Tau3 Banking](https://benchgap.net/calibration/aa-agentic-index/current/to/aa-tau3-banking/current.md) | linear | 15 | 3.7 pp | 31 |
| [SWE-bench Pro → PostTrainBench](https://benchgap.net/calibration/swe-bench-pro/current/to/post-train-bench12/current.md) | inverse Michaelis–Menten | 9 | 3.7 pp | 39 |
| [Vals SWE-bench → PostTrainBench v1.1](https://benchgap.net/calibration/valsswebench/current/to/posttrainbench%23posttrainbench-v1-1/current.md) | Hill | 10 | 3.7 pp | 44 |
| [AA-MMMU-Pro → Video-MME (with subtitle)](https://benchgap.net/calibration/aa-mmmu-pro/current/to/videommewithsub/current.md) | Michaelis–Menten | 8 | 3.7 pp | 40 |
| [INCLUDE → SWE Multilingual](https://benchgap.net/calibration/include/current/to/swe-bench-multilingual/current.md) | offset logistic | 7 | 3.7 pp | 1 |
| [AA-SciCode → cursorBench41](https://benchgap.net/calibration/aa-sci-code/current/to/cursor-bench41/current.md) | offset logistic | 9 | 3.7 pp | 61 |
| [ParseBench-2 → ParseBench-1](https://benchgap.net/calibration/parsebench/current/to/parse-bench-llama-index/current.md) | linear | 16 | 3.7 pp | 5 |
| [DeepPlanning → QwenClawBench](https://benchgap.net/calibration/deepplanning/current/to/qwenclawbench/current.md) | linear | 7 | 3.7 pp | 0 |
| [FrontierSWE v2 → AA-SciCode](https://benchgap.net/calibration/frontierswev2/current/to/aa-sci-code/current.md) | Michaelis–Menten | 20 | 3.7 pp | 0 |
| [Hermes Index → MKA](https://benchgap.net/calibration/hermesindex/current/to/mka/current.md) | Michaelis–Menten | 10 | 3.7 pp | 2 |
| [SuperGPQA → AA-GPQA Diamond](https://benchgap.net/calibration/supergpqa/current/to/aa-gpqa-diamond/current.md) | offset logistic | 17 | 3.7 pp | 4 |
| [MCP Atlas → AA-AnalystAgent](https://benchgap.net/calibration/mcpatlas/current/to/aa-analyst-agent/current.md) | offset logistic | 6 | 3.7 pp | 29 |
| [CursorBench 4.0 → Vals SWE-bench](https://benchgap.net/calibration/cursor-bench40/current/to/valsswebench/current.md) | linear | 9 | 3.7 pp | 9 |
| [ExploitGym → MazeBench](https://benchgap.net/calibration/exploitgym/current/to/mazebench/current.md) | offset logistic | 15 | 3.7 pp | 8 |
| [CharXiv → MMMU](https://benchgap.net/calibration/charxiv/current/to/mmmu/current.md) | Hill | 13 | 3.7 pp | 8 |
| [AA-Omniscience Accuracy → BioBasicsBench](https://benchgap.net/calibration/omniscienceaccuracy/current/to/biobasicsbench/current.md) | offset logistic | 7 | 3.8 pp | 46 |
| [Vibe Code Bench → AA-SciCode](https://benchgap.net/calibration/vibecodebench/current/to/aa-sci-code/current.md) | linear | 11 | 3.8 pp | 17 |
| [MMMU → MMVU](https://benchgap.net/calibration/mmmu/current/to/mmvu/current.md) | Michaelis–Menten | 5 | 3.8 pp | 12 |
| [AA-GPQA Diamond → CritPt](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/critpt/current.md) | offset logistic | 176 | 3.8 pp | 3 |
| [CharXiv w/o tools → PerceptionBench](https://benchgap.net/calibration/charxivnotools/current/to/perceptionbench/current.md) | offset logistic | 6 | 3.8 pp | 8 |
| [ScreenSpot Pro → CharXiv](https://benchgap.net/calibration/screenspot-pro/current/to/charxiv/current.md) | Michaelis–Menten | 14 | 3.8 pp | 6 |
| [DeepSWE → ScarfBench](https://benchgap.net/calibration/deepswe/current/to/scarfbench/current.md) | Hill | 9 | 3.8 pp | 56 |
| [SimpleQA → MMLU](https://benchgap.net/calibration/simpleqa/current/to/mmlu/current.md) | Michaelis–Menten | 7 | 3.8 pp | 1 |
| [HMMT Feb 2026 → AIME 2025](https://benchgap.net/calibration/hmmtfeb2026/current/to/aime2025/current.md) | offset logistic | 5 | 3.8 pp | 5 |
| [MazeBench → LifeVal](https://benchgap.net/calibration/mazebench/current/to/lifeval/current.md) | offset logistic | 6 | 3.8 pp | 26 |
| [MazeBench → Finance Agent v2](https://benchgap.net/calibration/mazebench/current/to/financeagentv2/current.md) | offset logistic | 6 | 3.8 pp | 49 |
| [MLVU (M-Avg) → AA-MMMU-Pro](https://benchgap.net/calibration/mlvuavg/current/to/aa-mmmu-pro/current.md) | offset logistic | 7 | 3.8 pp | 0 |
| [Artificial Analysis Intelligence Index → MMLU](https://benchgap.net/calibration/artificialanalysis/current/to/mmlu/current.md) | offset logistic | 14 | 3.8 pp | 13 |
| [RealWorldQA → CharXiv w/o tools](https://benchgap.net/calibration/realworldqa/current/to/charxivnotools/current.md) | Hill | 5 | 3.8 pp | 1 |
| [AA-GPQA Diamond → MMLU-Redux](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/mmluredux/current.md) | linear | 9 | 3.8 pp | 0 |
| [ProgramBench → DeepSWE](https://benchgap.net/calibration/programbench/current/to/deepswe/current.md) | offset logistic | 19 | 3.8 pp | 2 |
| [MathVision → CharXiv](https://benchgap.net/calibration/mathvision/current/to/charxiv/current.md) | linear | 22 | 3.8 pp | 1 |
| [FrontierSWE v2 → FrontierCode 1.1 Main](https://benchgap.net/calibration/frontierswev2/current/to/frontiercode/current.md) | Michaelis–Menten | 7 | 3.8 pp | 3 |
| [MMMU-Pro → BabyVision w/ Python](https://benchgap.net/calibration/mmmu-pro/current/to/babyvisionpython/current.md) | offset logistic | 5 | 3.8 pp | 40 |
| [HLE-Verified → AA-HLE](https://benchgap.net/calibration/hleverified/current/to/aa-hle/current.md) | offset logistic | 6 | 3.8 pp | 0 |
| [FrontierCode 1.1 Main → AA-SciCode](https://benchgap.net/calibration/frontiercode/current/to/aa-sci-code/current.md) | linear | 11 | 3.8 pp | 0 |
| [SimpleVQA → RealWorldQA](https://benchgap.net/calibration/simplevqa/current/to/realworldqa/current.md) | Michaelis–Menten | 11 | 3.8 pp | 0 |
| [SWE-bench Verified → AA-SciCode](https://benchgap.net/calibration/swe-bench-verified/current/to/aa-sci-code/current.md) | Michaelis–Menten | 35 | 3.8 pp | 42 |
| [SimpleVQA → CountBench](https://benchgap.net/calibration/simplevqa/current/to/countbench/current.md) | linear | 8 | 3.8 pp | 0 |
| [MKA → CWE-Bench-AA](https://benchgap.net/calibration/mka/current/to/aa-cwe-bench/current.md) | offset logistic | 11 | 3.8 pp | 11 |
| [Claw-Eval → OSWorld-Verified](https://benchgap.net/calibration/claw-eval/current/to/osworld-verified/current.md) | Hill | 9 | 3.8 pp | 32 |
| [ExploitBench → SEC-Bench Pro](https://benchgap.net/calibration/exploitbench/current/to/secbenchpro/current.md) | Michaelis–Menten | 9 | 3.9 pp | 9 |
| [DeepSWE → AA-SciCode](https://benchgap.net/calibration/deepswe/current/to/aa-sci-code/current.md) | Michaelis–Menten | 40 | 3.9 pp | 9 |
| [PostTrainBench → Vals SWE-bench](https://benchgap.net/calibration/post-train-bench12/current/to/valsswebench/current.md) | linear | 12 | 3.9 pp | 6 |
| [FrontierCode 2.0 → FrontierCode 1.1 Main](https://benchgap.net/calibration/frontiercode-2/current/to/frontiercode/current.md) | linear | 10 | 3.9 pp | 2 |
| [FrontierPhysics → DAYJOB](https://benchgap.net/calibration/frontierphysics/current/to/dayjob/current.md) | inverse Michaelis–Menten | 12 | 3.9 pp | 1 |
| [VideoMMMU → Video-MME (with subtitle)](https://benchgap.net/calibration/videommmu/current/to/videommewithsub/current.md) | Michaelis–Menten + offset | 11 | 3.9 pp | 0 |
| [GPQA Diamond → AA-HLE](https://benchgap.net/calibration/gpqa-diamond/current/to/aa-hle/current.md) | inverse Michaelis–Menten | 50 | 3.9 pp | 36 |
| [Terminal-Bench 2.1 (Vals AI) → Terminal-Bench Hard](https://benchgap.net/calibration/vals-terminal-bench21/current/to/terminal-bench-hard/current.md) | Hill | 13 | 3.9 pp | 27 |
| [HLE w/o tools → ARC-AGI-1](https://benchgap.net/calibration/hlenotools/current/to/arcagi1/current.md) | Hill | 14 | 3.9 pp | 13 |
| [Toolathlon-Verified → WideResearch](https://benchgap.net/calibration/toolathlonverified/current/to/wideresearch/current.md) | Michaelis–Menten | 6 | 3.9 pp | 5 |
| [Bug Hunt Bench → VulcanBench v3](https://benchgap.net/calibration/bug-hunt-bench/current/to/vulcanbench%23vulcanbench-v3/current.md) | offset logistic | 7 | 3.9 pp | 20 |
| [AA Agentic Index → AA ITBench](https://benchgap.net/calibration/aa-agentic-index/current/to/aa-itbench/current.md) | linear | 17 | 4.0 pp | 40 |
| [MCP Atlas → DRACO](https://benchgap.net/calibration/mcpatlas/current/to/draco/current.md) | Michaelis–Menten | 6 | 4.0 pp | 4 |
| [Terminal-Bench 3.0 → Terminal-Bench 2.1 (Vals AI)](https://benchgap.net/calibration/terminal-bench%23terminal-bench-3-0/current/to/vals-terminal-bench21/current.md) | linear | 11 | 4.0 pp | 0 |
| [MMLU → MMLU-Pro](https://benchgap.net/calibration/mmlu/current/to/mmlu-pro/current.md) | inverse Michaelis–Menten | 19 | 4.0 pp | 13 |
| [We-Math → Video-MME (with subtitle)](https://benchgap.net/calibration/wemath/current/to/videommewithsub/current.md) | Hill | 6 | 4.0 pp | 0 |
| [AA-MMMU-Pro → BabyVision w/ Python](https://benchgap.net/calibration/aa-mmmu-pro/current/to/babyvisionpython/current.md) | Michaelis–Menten | 6 | 4.0 pp | 33 |
| [Artificial Analysis Decision Index → Artificial Analysis Decision Index v1.0](https://benchgap.net/calibration/aa-decision-index-v2/current/to/aa-decision-index/current.md) | Hill | 14 | 4.0 pp | 5 |
| [CritPt → MMLU-Pro (Arcee)](https://benchgap.net/calibration/critpt/current/to/mmlu-pro-arcee/current.md) | linear | 6 | 4.0 pp | 107 |
| [DeepsecBench-AA → BrowseComp](https://benchgap.net/calibration/aa-deepsec-bench/current/to/browsecomp/current.md) | Michaelis–Menten | 8 | 4.0 pp | 3 |
| [CyberGym → ApprenticeBench](https://benchgap.net/calibration/cybergym/current/to/apprenticebench/current.md) | offset logistic | 8 | 4.0 pp | 17 |
| [AIME26 → HMMT Feb 2026](https://benchgap.net/calibration/aime2026/current/to/hmmtfeb2026/current.md) | inverse Michaelis–Menten | 22 | 4.0 pp | 11 |
| [AA Coding Index → VulcanBench Frontier v4](https://benchgap.net/calibration/aa-coding-index/current/to/vulcanbench/current.md) | Michaelis–Menten + offset | 6 | 4.0 pp | 10 |
| [NYT Connections → FACTS Multimodal](https://benchgap.net/calibration/nyt-connections/current/to/facts-multimodal/current.md) | linear | 22 | 4.0 pp | 52 |
| [GPQA → CritPt](https://benchgap.net/calibration/gpqa/current/to/critpt/current.md) | offset logistic | 84 | 4.0 pp | 47 |
| [Terminal-Bench 2.0 → Terminal-Bench Hard](https://benchgap.net/calibration/terminal-bench%23terminal-bench-2-0/current/to/terminal-bench-hard/current.md) | Hill | 12 | 4.0 pp | 16 |
| [AA-AnalystAgent → Harvey LAB-AA v1.1](https://benchgap.net/calibration/aa-analyst-agent/current/to/aa-harvey-lab-v11/current.md) | offset logistic | 6 | 4.0 pp | 12 |
| [Toolathlon-Verified → MazeBench](https://benchgap.net/calibration/toolathlonverified/current/to/mazebench/current.md) | linear | 10 | 4.0 pp | 6 |
| [CWE-bench → CWE-Bench-AA](https://benchgap.net/calibration/cwe-bench-v2/current/to/aa-cwe-bench/current.md) | Hill | 9 | 4.0 pp | 3 |
| [Chess Puzzles → BlenderWorld (text-only)](https://benchgap.net/calibration/chesspuzzles/current/to/blenderworld-text-only/current.md) | offset logistic | 5 | 4.0 pp | 19 |
| [MLVU (M-Avg) → Video-MME (w/o subtitle)](https://benchgap.net/calibration/mlvuavg/current/to/videommenosub/current.md) | inverse Michaelis–Menten | 7 | 4.0 pp | 0 |
| [FrontierMath v2 (Tiers 1-3) → FrontierMath v2 (Tier 4)](https://benchgap.net/calibration/frontiermathv2tiers13/current/to/frontiermathv2tier4/current.md) | inverse Michaelis–Menten | 58 | 4.0 pp | 13 |
| [DeepsecBench-AA → Hermes Bench](https://benchgap.net/calibration/aa-deepsec-bench/current/to/hermesbench/current.md) | Michaelis–Menten | 10 | 4.0 pp | 0 |
| [Furniture Assembly → DAYJOB Healthcare](https://benchgap.net/calibration/furnitureassembly/current/to/dayjob-healthcare/current.md) | linear | 16 | 4.0 pp | 19 |
| [MMLU-Pro → CritPt](https://benchgap.net/calibration/mmlu-pro/current/to/critpt/current.md) | inverse Michaelis–Menten | 49 | 4.1 pp | 4 |
| [FrontierCode 2.0 → CursorBench 3.2](https://benchgap.net/calibration/frontiercode-2/current/to/cursor-bench32/current.md) | Hill | 6 | 4.1 pp | 7 |
| [GDPval-AA → AA Tau3 Banking](https://benchgap.net/calibration/gdpvalaanormalized/current/to/aa-tau3-banking/current.md) | Michaelis–Menten | 17 | 4.1 pp | 38 |
| [BrowseComp → SkillsBench](https://benchgap.net/calibration/browsecomp/current/to/skills-bench/current.md) | inverse Michaelis–Menten | 5 | 4.1 pp | 52 |
| [Cookie-Bench → BlenderWorld (text-only)](https://benchgap.net/calibration/cookie-bench/current/to/blenderworld-text-only/current.md) | inverse Michaelis–Menten | 7 | 4.1 pp | 0 |
| [PostTrainBench v1.1 → Vals SWE-bench](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1-1/current/to/valsswebench/current.md) | Hill | 10 | 4.1 pp | 1 |
| [LABBench2 → AA-HLE](https://benchgap.net/calibration/labbench2/current/to/aa-hle/current.md) | offset logistic | 7 | 4.1 pp | 0 |
| [HMMT Feb 2026 → FrontierMath v2 (Tier 4)](https://benchgap.net/calibration/hmmtfeb2026/current/to/frontiermathv2tier4/current.md) | linear | 6 | 4.1 pp | 11 |
| [AA-MMMU-Pro → Chartography (tools)](https://benchgap.net/calibration/aa-mmmu-pro/current/to/chartographywithtools/current.md) | Michaelis–Menten | 5 | 4.1 pp | 70 |
| [FrontierSWE v2 → CursorBench 3.2](https://benchgap.net/calibration/frontierswev2/current/to/cursor-bench32/current.md) | Hill | 7 | 4.1 pp | 0 |
| [EBR-bench → DAYJOB](https://benchgap.net/calibration/ebrbench/current/to/dayjob/current.md) | Hill | 11 | 4.1 pp | 0 |
| [MCP-Tasks → DeepPlanning](https://benchgap.net/calibration/mcptasks/current/to/deepplanning/current.md) | inverse Michaelis–Menten | 5 | 4.1 pp | 0 |
| [HealthBench (length-adjusted) → HealthBench Professional (raw)](https://benchgap.net/calibration/healthbenchlengthadjusted/current/to/healthbenchprofessionalraw/current.md) | Hill | 8 | 4.1 pp | 0 |
| [BioBasicsBench → ARC-AGI-1](https://benchgap.net/calibration/biobasicsbench/current/to/arcagi1/current.md) | Michaelis–Menten | 7 | 4.1 pp | 0 |
| [CursorBench 4.0 → DeepSWE](https://benchgap.net/calibration/cursor-bench40/current/to/deepswe/current.md) | Michaelis–Menten + offset | 16 | 4.1 pp | 5 |
| [JobBench → GDP.pdf](https://benchgap.net/calibration/jobbench/current/to/aa-gdp-pdf/current.md) | offset logistic | 5 | 4.1 pp | 31 |
| [BrowseComp → AA ITBench](https://benchgap.net/calibration/browsecomp/current/to/aa-itbench/current.md) | Michaelis–Menten | 10 | 4.1 pp | 31 |
| [Mystery Game Puzzles → OTelBench](https://benchgap.net/calibration/mysterygamepuzzles/current/to/otelbench/current.md) | Michaelis–Menten | 5 | 4.1 pp | 26 |
| [Chartography (tools) → AA-MMMU-Pro](https://benchgap.net/calibration/chartographywithtools/current/to/aa-mmmu-pro/current.md) | Michaelis–Menten | 5 | 4.1 pp | 0 |
| [ScarfBench → PostTrainBench](https://benchgap.net/calibration/scarfbench/current/to/post-train-bench12/current.md) | linear | 7 | 4.1 pp | 1 |
| [MathVision → DynaMath](https://benchgap.net/calibration/mathvision/current/to/dynamath/current.md) | linear | 6 | 4.1 pp | 1 |
| [APEX-Agents-AA → HLE w/ tools](https://benchgap.net/calibration/apexagentsaa/current/to/hlewithtools/current.md) | Hill | 5 | 4.1 pp | 21 |
| [MMLU-Redux → SuperGPQA](https://benchgap.net/calibration/mmluredux/current/to/supergpqa/current.md) | Hill | 9 | 4.1 pp | 6 |
| [CursorBench 4.0 → SWE-bench Pro](https://benchgap.net/calibration/cursor-bench40/current/to/swe-bench-pro/current.md) | Hill | 9 | 4.2 pp | 8 |
| [CountBench → RealWorldQA](https://benchgap.net/calibration/countbench/current/to/realworldqa/current.md) | inverse Michaelis–Menten | 10 | 4.2 pp | 0 |
| [RealWorldQA → MLVU (M-Avg)](https://benchgap.net/calibration/realworldqa/current/to/mlvuavg/current.md) | inverse Michaelis–Menten | 10 | 4.2 pp | 1 |
| [BBH → GPQA](https://benchgap.net/calibration/bbh/current/to/gpqa/current.md) | linear | 7 | 4.2 pp | 0 |
| [OSWorld 2.1 → GDPval-AA](https://benchgap.net/calibration/osworld-2-1/current/to/gdpvalaanormalized/current.md) | linear | 9 | 4.2 pp | 3 |
| [Artificial Analysis Intelligence Index → Vals MMLU-Pro](https://benchgap.net/calibration/artificialanalysis/current/to/vals-mmlu-pro/current.md) | Michaelis–Menten | 72 | 4.2 pp | 2 |
| [RealWorldQA → ERQA](https://benchgap.net/calibration/realworldqa/current/to/erqa/current.md) | Hill | 11 | 4.2 pp | 3 |
| [HLE w/o tools → HealthBench Professional](https://benchgap.net/calibration/hlenotools/current/to/healthbenchprofessional/current.md) | Michaelis–Menten + offset | 6 | 4.2 pp | 38 |
| [ERQA → MedXpertQA (MM)](https://benchgap.net/calibration/erqa/current/to/medxpertqamm/current.md) | Hill | 7 | 4.2 pp | 5 |
| [GPQA → HealthBench Professional](https://benchgap.net/calibration/gpqa/current/to/healthbenchprofessional/current.md) | Hill | 7 | 4.2 pp | 15 |
| [GPQA Diamond → HealthBench Professional](https://benchgap.net/calibration/gpqa-diamond/current/to/healthbenchprofessional/current.md) | Hill | 7 | 4.2 pp | 6 |
| [Video-MME (with subtitle) → MMMU-Pro](https://benchgap.net/calibration/videommewithsub/current/to/mmmu-pro/current.md) | inverse Michaelis–Menten | 12 | 4.2 pp | 0 |
| [Vals SWE-bench → AA-SciCode](https://benchgap.net/calibration/valsswebench/current/to/aa-sci-code/current.md) | linear | 42 | 4.2 pp | 2 |
| [BrowseComp → Toolathlon](https://benchgap.net/calibration/browsecomp/current/to/toolathlon/current.md) | inverse Michaelis–Menten | 13 | 4.2 pp | 45 |
| [Artificial Analysis Intelligence Index → HealthBench Professional](https://benchgap.net/calibration/artificialanalysis/current/to/healthbenchprofessional/current.md) | Hill | 19 | 4.2 pp | 118 |
| [APEX-Agents-AA → BrowseComp](https://benchgap.net/calibration/apexagentsaa/current/to/browsecomp/current.md) | Michaelis–Menten + offset | 12 | 4.2 pp | 16 |
| [HLE-MultiChoice (Decision Index) → Artificial Analysis Decision Index v1.0](https://benchgap.net/calibration/aa-decision-hle-multi-choice/current/to/aa-decision-index/current.md) | Michaelis–Menten | 19 | 4.2 pp | 0 |
| [SWE-bench Verified → FrontierCode 1.1 Main](https://benchgap.net/calibration/swe-bench-verified/current/to/frontiercode/current.md) | linear | 7 | 4.2 pp | 67 |
| [Bug Hunt Bench → AA-SciCode](https://benchgap.net/calibration/bug-hunt-bench/current/to/aa-sci-code/current.md) | Michaelis–Menten | 23 | 4.2 pp | 0 |
| [GPQA → MMLU](https://benchgap.net/calibration/gpqa/current/to/mmlu/current.md) | Hill | 26 | 4.2 pp | 29 |
| [Mystery Game Puzzles → DAYJOB Healthcare](https://benchgap.net/calibration/mysterygamepuzzles/current/to/dayjob-healthcare/current.md) | offset logistic | 19 | 4.2 pp | 33 |
| [FrontierSWE v2 → cursorBench41](https://benchgap.net/calibration/frontierswev2/current/to/cursor-bench41/current.md) | linear | 12 | 4.3 pp | 1 |
| [FrontierSWE v2 → PostTrainBench v1.1](https://benchgap.net/calibration/frontierswev2/current/to/posttrainbench%23posttrainbench-v1-1/current.md) | linear | 7 | 4.3 pp | 1 |
| [PostTrainBench v1.1 → DeepSWE](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1-1/current/to/deepswe/current.md) | Michaelis–Menten | 9 | 4.3 pp | 5 |
| [ParseBench-1 → ParseBench-2](https://benchgap.net/calibration/parse-bench-llama-index/current/to/parsebench/current.md) | Michaelis–Menten | 16 | 4.3 pp | 8 |
| [GPQA Diamond → HLE w/o tools](https://benchgap.net/calibration/gpqa-diamond/current/to/hlenotools/current.md) | offset logistic | 36 | 4.3 pp | 32 |
| [ARC-AGI-2 → HealthBench Professional](https://benchgap.net/calibration/arc-agi-2/current/to/healthbenchprofessional/current.md) | Hill | 13 | 4.3 pp | 1 |
| [EBR-bench → MaxxBench](https://benchgap.net/calibration/ebrbench/current/to/maxxbench/current.md) | Michaelis–Menten | 5 | 4.3 pp | 25 |
| [Hermes Index → CWE-bench](https://benchgap.net/calibration/hermesindex/current/to/cwe-bench-v2/current.md) | Michaelis–Menten + offset | 7 | 4.3 pp | 9 |
| [Artificial Analysis Intelligence Index → HealthBench Professional (length-adjusted, Artificial Analysis)](https://benchgap.net/calibration/artificialanalysis/current/to/aa-health-bench-professional/current.md) | linear | 6 | 4.3 pp | 145 |
| [Video-MME (w/o subtitle) → CountBench](https://benchgap.net/calibration/videommenosub/current/to/countbench/current.md) | offset logistic | 6 | 4.3 pp | 0 |
| [MLVU (M-Avg) → CharXiv](https://benchgap.net/calibration/mlvuavg/current/to/charxiv/current.md) | Hill | 9 | 4.3 pp | 0 |
| [NL2Repo → AA Coding Index](https://benchgap.net/calibration/nl2repo/current/to/aa-coding-index/current.md) | Hill | 15 | 4.3 pp | 11 |
| [PostTrainBench → SWE-bench Pro](https://benchgap.net/calibration/post-train-bench12/current/to/swe-bench-pro/current.md) | Hill | 9 | 4.3 pp | 8 |
| [FrontierSWE v2 → DeepSWE](https://benchgap.net/calibration/frontierswev2/current/to/deepswe/current.md) | Michaelis–Menten + offset | 19 | 4.3 pp | 3 |
| [NYT Connections → BlenderWorld (text-only)](https://benchgap.net/calibration/nyt-connections/current/to/blenderworld-text-only/current.md) | Hill | 5 | 4.3 pp | 10 |
| [ExploitBench Internal Port → FrontierPhysics](https://benchgap.net/calibration/exploitbenchinternalport/current/to/frontierphysics/current.md) | linear | 7 | 4.3 pp | 1 |
| [OfficeQA Pro → Chartography (tools)](https://benchgap.net/calibration/officeqapro/current/to/chartographywithtools/current.md) | Michaelis–Menten | 7 | 4.3 pp | 3 |
| [CritPt → Vals MMLU-Pro](https://benchgap.net/calibration/critpt/current/to/vals-mmlu-pro/current.md) | Hill | 70 | 4.3 pp | 1 |
| [AIME 2025 → AIME26](https://benchgap.net/calibration/aime2025/current/to/aime2026/current.md) | linear | 7 | 4.3 pp | 35 |
| [FrontierCode 1.1 Main → PostTrainBench](https://benchgap.net/calibration/frontiercode/current/to/post-train-bench12/current.md) | inverse Michaelis–Menten | 6 | 4.3 pp | 3 |
| [Vibe Code Bench → FrontierCode 1.1 Main](https://benchgap.net/calibration/vibecodebench/current/to/frontiercode/current.md) | inverse Michaelis–Menten | 5 | 4.3 pp | 5 |
| [Hermes Index → HLE w/ tools](https://benchgap.net/calibration/hermesindex/current/to/hlewithtools/current.md) | linear | 8 | 4.3 pp | 1 |
| [AA-Omniscience Accuracy → SimpleQA](https://benchgap.net/calibration/omniscienceaccuracy/current/to/simpleqa/current.md) | linear | 7 | 4.3 pp | 0 |
| [VulcanBench Frontier v4 → AA-SciCode](https://benchgap.net/calibration/vulcanbench/current/to/aa-sci-code/current.md) | Michaelis–Menten + offset | 18 | 4.3 pp | 0 |
| [Hermes Bench → AA Agentic Index](https://benchgap.net/calibration/hermesbench/current/to/aa-agentic-index/current.md) | Hill | 5 | 4.3 pp | 0 |
| [Toolathlon → BrowseComp](https://benchgap.net/calibration/toolathlon/current/to/browsecomp/current.md) | linear | 13 | 4.3 pp | 3 |
| [DynaMath → MLVU (M-Avg)](https://benchgap.net/calibration/dynamath/current/to/mlvuavg/current.md) | Michaelis–Menten + offset | 7 | 4.3 pp | 0 |
| [Hermes Index → DRACO](https://benchgap.net/calibration/hermesindex/current/to/draco/current.md) | Michaelis–Menten | 5 | 4.3 pp | 0 |
| [PhysicianBench → HealthBench (raw, Artificial Analysis)](https://benchgap.net/calibration/physicianbench/current/to/aa-health-bench/current.md) | Michaelis–Menten + offset | 5 | 4.3 pp | 0 |
| [ARC-AGI-3 → ARC-AGI-3 Provider Adapter](https://benchgap.net/calibration/arcagi3/current/to/arc-agi-3-provider-adapter/current.md) | Michaelis–Menten | 13 | 4.4 pp | 4 |
| [HLE w/o tools → HealthBench Professional (length-adjusted, Artificial Analysis)](https://benchgap.net/calibration/hlenotools/current/to/aa-health-bench-professional/current.md) | Hill | 5 | 4.4 pp | 20 |
| [Vals SWE-bench → SWE-bench Verified](https://benchgap.net/calibration/valsswebench/current/to/swe-bench-verified/current.md) | linear | 26 | 4.4 pp | 21 |
| [AA Coding Index → CursorBench 3.2](https://benchgap.net/calibration/aa-coding-index/current/to/cursor-bench32/current.md) | inverse Michaelis–Menten | 17 | 4.4 pp | 17 |
| [Video-MME (w/o subtitle) → MLVU (M-Avg)](https://benchgap.net/calibration/videommenosub/current/to/mlvuavg/current.md) | Hill | 7 | 4.4 pp | 0 |
| [SWE-bench Pro → PostTrainBench v1.1](https://benchgap.net/calibration/swe-bench-pro/current/to/posttrainbench%23posttrainbench-v1-1/current.md) | inverse Michaelis–Menten | 10 | 4.4 pp | 39 |
| [Furniture Assembly → FrontierPhysics](https://benchgap.net/calibration/furnitureassembly/current/to/frontierphysics/current.md) | inverse Michaelis–Menten | 24 | 4.4 pp | 15 |
| [HealthBench (length-adjusted) → ARC-AGI-1](https://benchgap.net/calibration/healthbenchlengthadjusted/current/to/arcagi1/current.md) | Michaelis–Menten | 6 | 4.4 pp | 1 |
| [DeepsecBench-AA → GDP.pdf](https://benchgap.net/calibration/aa-deepsec-bench/current/to/aa-gdp-pdf/current.md) | linear | 11 | 4.4 pp | 3 |
| [LifeVal → MazeBench](https://benchgap.net/calibration/lifeval/current/to/mazebench/current.md) | offset logistic | 6 | 4.4 pp | 0 |
| [ApprenticeBench → AA ITBench](https://benchgap.net/calibration/apprenticebench/current/to/aa-itbench/current.md) | Michaelis–Menten + offset | 9 | 4.4 pp | 0 |
| [Decision Index Full Score → Decision Index Language](https://benchgap.net/calibration/decisionindex/current/to/jevdecisionindex03language/current.md) | Hill | 110 | 4.4 pp | 0 |
| [ARC-AGI-1 → HealthBench Professional](https://benchgap.net/calibration/arcagi1/current/to/healthbenchprofessional/current.md) | offset logistic | 13 | 4.4 pp | 0 |
| [CountBench → CC-OCR](https://benchgap.net/calibration/countbench/current/to/ccocr/current.md) | Michaelis–Menten | 5 | 4.4 pp | 0 |
| [GPQA → MMLU-Pro](https://benchgap.net/calibration/gpqa/current/to/mmlu-pro/current.md) | Hill | 62 | 4.4 pp | 35 |
| [C-Eval → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/ceval/current/to/artificialanalysis/current.md) | offset logistic | 5 | 4.4 pp | 4 |
| [MKA → CWE-bench](https://benchgap.net/calibration/mka/current/to/cwe-bench-v2/current.md) | linear | 14 | 4.4 pp | 2 |
| [AI2D_TEST → Video-MME (w/o subtitle)](https://benchgap.net/calibration/ai2dtest/current/to/videommenosub/current.md) | Michaelis–Menten + offset | 7 | 4.4 pp | 0 |
| [GPQA Diamond → AA-GPQA Diamond](https://benchgap.net/calibration/gpqa-diamond/current/to/aa-gpqa-diamond/current.md) | Hill | 41 | 4.4 pp | 23 |
| [MCP Atlas → AA ITBench](https://benchgap.net/calibration/mcpatlas/current/to/aa-itbench/current.md) | Michaelis–Menten + offset | 9 | 4.4 pp | 6 |
| [SciCode → LiveCodeBench v6](https://benchgap.net/calibration/scicode/current/to/livecodebench-v6/current.md) | Hill | 14 | 4.4 pp | 9 |
| [FrontierCode 1.1 Main → AA Coding Index](https://benchgap.net/calibration/frontiercode/current/to/aa-coding-index/current.md) | Michaelis–Menten | 9 | 4.4 pp | 1 |
| [BrowseComp → APEX-Agents-AA](https://benchgap.net/calibration/browsecomp/current/to/apexagentsaa/current.md) | Hill | 12 | 4.4 pp | 37 |
| [SWE-bench Verified → SWE-bench Pro](https://benchgap.net/calibration/swe-bench-verified/current/to/swe-bench-pro/current.md) | linear | 51 | 4.5 pp | 60 |
| [GPQA → Vals GPQA Diamond](https://benchgap.net/calibration/gpqa/current/to/vals-gpqa-diamond/current.md) | linear | 41 | 4.5 pp | 34 |
| [Vals GPQA Diamond → GPQA](https://benchgap.net/calibration/vals-gpqa-diamond/current/to/gpqa/current.md) | linear | 41 | 4.5 pp | 24 |
| [MMVU → MMMU-Pro](https://benchgap.net/calibration/mmvu/current/to/mmmu-pro/current.md) | inverse Michaelis–Menten | 7 | 4.5 pp | 0 |
| [Terminal-Bench 4.1 → Terminal-Bench 4.0](https://benchgap.net/calibration/terminal-bench41/current/to/terminal-bench/current.md) | linear | 17 | 4.5 pp | 23 |
| [AA-HLE → Vals MMLU-Pro](https://benchgap.net/calibration/aa-hle/current/to/vals-mmlu-pro/current.md) | Michaelis–Menten | 74 | 4.5 pp | 0 |
| [GPQA Diamond → MedXpertQA (Text)](https://benchgap.net/calibration/gpqa-diamond/current/to/medxpertqatext/current.md) | inverse Michaelis–Menten | 5 | 4.5 pp | 60 |
| [DeepSWE → PostTrainBench v1.1](https://benchgap.net/calibration/deepswe/current/to/posttrainbench%23posttrainbench-v1-1/current.md) | inverse Michaelis–Menten | 9 | 4.5 pp | 21 |
| [Vals SWE-bench → SciCode](https://benchgap.net/calibration/valsswebench/current/to/scicode/current.md) | Michaelis–Menten | 8 | 4.5 pp | 3 |
| [LifeVal → AA-AnalystAgent](https://benchgap.net/calibration/lifeval/current/to/aa-analyst-agent/current.md) | inverse Michaelis–Menten | 5 | 4.5 pp | 0 |
| [Artificial Analysis Decision Index → Decision Index Arts](https://benchgap.net/calibration/aa-decision-index-v2/current/to/jevdecisionindex03arts/current.md) | linear | 11 | 4.5 pp | 5 |
| [Video-MME (w/o subtitle) → MathVision](https://benchgap.net/calibration/videommenosub/current/to/mathvision/current.md) | linear | 6 | 4.5 pp | 0 |
| [JobBench → Claw-Eval](https://benchgap.net/calibration/jobbench/current/to/claw-eval/current.md) | linear | 7 | 4.5 pp | 28 |
| [MCP Atlas → AA Tau3 Banking](https://benchgap.net/calibration/mcpatlas/current/to/aa-tau3-banking/current.md) | Michaelis–Menten | 7 | 4.5 pp | 2 |
| [FrontierCode 2.0 → PostTrainBench](https://benchgap.net/calibration/frontiercode-2/current/to/post-train-bench12/current.md) | Hill | 15 | 4.5 pp | 3 |
| [ApprenticeBench → CyberGym](https://benchgap.net/calibration/apprenticebench/current/to/cybergym/current.md) | linear | 8 | 4.5 pp | 3 |
| [FrontierSWE v2 → ScarfBench](https://benchgap.net/calibration/frontierswev2/current/to/scarfbench/current.md) | Michaelis–Menten | 7 | 4.5 pp | 6 |
| [Vibe Code Bench → SWE-bench Verified](https://benchgap.net/calibration/vibecodebench/current/to/swe-bench-verified/current.md) | linear | 11 | 4.5 pp | 9 |
| [Video-MME (with subtitle) → MathVision](https://benchgap.net/calibration/videommewithsub/current/to/mathvision/current.md) | linear | 8 | 4.5 pp | 0 |
| [OSWorld 2.0 → CWE-Bench-AA](https://benchgap.net/calibration/osworld2/current/to/aa-cwe-bench/current.md) | linear | 7 | 4.5 pp | 14 |
| [AA-HLE → HealthBench Professional](https://benchgap.net/calibration/aa-hle/current/to/healthbenchprofessional/current.md) | offset logistic | 17 | 4.5 pp | 12 |
| [CursorBench 3.2 → SWE-bench Pro](https://benchgap.net/calibration/cursor-bench32/current/to/swe-bench-pro/current.md) | Hill | 12 | 4.5 pp | 3 |
| [AA Agentic Index → DeepSearchQA](https://benchgap.net/calibration/aa-agentic-index/current/to/deepsearchqa/current.md) | Michaelis–Menten | 8 | 4.5 pp | 61 |
| [FrontierCode 1.1 Main → DeepSWE](https://benchgap.net/calibration/frontiercode/current/to/deepswe/current.md) | Michaelis–Menten | 7 | 4.5 pp | 1 |
| [Vals LiveCodeBench → CursorBench 4.0](https://benchgap.net/calibration/valslivecodebench/current/to/cursor-bench40/current.md) | inverse Michaelis–Menten | 8 | 4.5 pp | 10 |
| [GPQA → BBH](https://benchgap.net/calibration/gpqa/current/to/bbh/current.md) | offset logistic | 7 | 4.5 pp | 9 |
| [HealthBench Professional (raw, Artificial Analysis) → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/aa-health-bench-professional-raw/current/to/artificialanalysis/current.md) | linear | 6 | 4.5 pp | 0 |
| [CursorBench 3.2 → CursorBench 3.1](https://benchgap.net/calibration/cursor-bench32/current/to/cursor-bench31/current.md) | Michaelis–Menten | 5 | 4.5 pp | 0 |
| [ExploitGym → CWE-bench v1](https://benchgap.net/calibration/exploitgym/current/to/cwe-bench%23cwe-bench-v1/current.md) | linear | 6 | 4.5 pp | 17 |
| [AA-HLE → HealthBench Professional (raw)](https://benchgap.net/calibration/aa-hle/current/to/healthbenchprofessionalraw/current.md) | Michaelis–Menten + offset | 8 | 4.5 pp | 103 |
| [HealthBench Hard → AA-GPQA Diamond](https://benchgap.net/calibration/healthbench-hard/current/to/aa-gpqa-diamond/current.md) | Hill | 8 | 4.5 pp | 0 |
| [OmniDocBench 1.5 → MathVision](https://benchgap.net/calibration/omnidocbench15/current/to/mathvision/current.md) | Hill | 8 | 4.5 pp | 0 |
| [MMMLU → GPQA](https://benchgap.net/calibration/mmmlu/current/to/gpqa/current.md) | offset logistic | 14 | 4.6 pp | 0 |
| [DAYJOB → FrontierPhysics](https://benchgap.net/calibration/dayjob/current/to/frontierphysics/current.md) | inverse Michaelis–Menten | 12 | 4.6 pp | 3 |
| [OSWorld 2.0 → Gert Labs](https://benchgap.net/calibration/osworld2/current/to/gertlabs/current.md) | Michaelis–Menten | 5 | 4.6 pp | 5 |
| [SWE-bench Pro → PostTrainBench v1](https://benchgap.net/calibration/swe-bench-pro/current/to/posttrainbench%23posttrainbench-v1/current.md) | linear | 9 | 4.6 pp | 36 |
| [RefCOCO (avg) → CountBench](https://benchgap.net/calibration/refcocoavg/current/to/countbench/current.md) | offset logistic | 7 | 4.6 pp | 0 |
| [Hermes Bench → CWE-Bench-AA](https://benchgap.net/calibration/hermesbench/current/to/aa-cwe-bench/current.md) | Michaelis–Menten | 10 | 4.6 pp | 1 |
| [Vals GPQA Diamond → HLE-Verified](https://benchgap.net/calibration/vals-gpqa-diamond/current/to/hleverified/current.md) | offset logistic | 6 | 4.6 pp | 15 |
| [DRACO → GDPval-AA](https://benchgap.net/calibration/draco/current/to/gdpvalaanormalized/current.md) | Hill | 7 | 4.6 pp | 2 |
| [Decision Index Tools → Artificial Analysis Decision Index v1.0](https://benchgap.net/calibration/jevdecisionindex03tools/current/to/aa-decision-index/current.md) | linear | 11 | 4.6 pp | 2 |
| [DynaMath → We-Math](https://benchgap.net/calibration/dynamath/current/to/wemath/current.md) | Hill | 6 | 4.6 pp | 0 |
| [AA-AnalystAgent → CWE-bench](https://benchgap.net/calibration/aa-analyst-agent/current/to/cwe-bench-v2/current.md) | Hill | 9 | 4.6 pp | 12 |
| [GDPval-AA → GDP.pdf](https://benchgap.net/calibration/gdpvalaanormalized/current/to/aa-gdp-pdf/current.md) | offset logistic | 13 | 4.6 pp | 97 |
| [Toolathlon-Verified → CWE-bench](https://benchgap.net/calibration/toolathlonverified/current/to/cwe-bench-v2/current.md) | offset logistic | 5 | 4.6 pp | 7 |
| [AA Coding Index → FrontierCode 1.1 Main](https://benchgap.net/calibration/aa-coding-index/current/to/frontiercode/current.md) | inverse Michaelis–Menten | 9 | 4.6 pp | 15 |
| [MKA → Hermes Index](https://benchgap.net/calibration/mka/current/to/hermesindex/current.md) | Hill | 10 | 4.6 pp | 1 |
| [GPQA Diamond → CritPt](https://benchgap.net/calibration/gpqa-diamond/current/to/critpt/current.md) | inverse Michaelis–Menten | 52 | 4.6 pp | 8 |
| [AutomationBench → CyberGym](https://benchgap.net/calibration/automationbench/current/to/cybergym/current.md) | Michaelis–Menten | 11 | 4.6 pp | 16 |
| [AA Coding Index → SciCode](https://benchgap.net/calibration/aa-coding-index/current/to/scicode/current.md) | linear | 15 | 4.6 pp | 11 |
| [Terminal-Bench 4.0 → Terminal-Bench 4.1](https://benchgap.net/calibration/terminal-bench/current/to/terminal-bench41/current.md) | Hill | 17 | 4.6 pp | 13 |
| [RealWorldQA → Video-MME (with subtitle)](https://benchgap.net/calibration/realworldqa/current/to/videommewithsub/current.md) | inverse Michaelis–Menten | 10 | 4.6 pp | 1 |
| [SWE-bench Multimodal → ParseBench-1](https://benchgap.net/calibration/swe-bench-multimodal/current/to/parse-bench-llama-index/current.md) | Hill | 5 | 4.6 pp | 2 |
| [HealthBench Professional (raw) → AA-HLE](https://benchgap.net/calibration/healthbenchprofessionalraw/current/to/aa-hle/current.md) | Michaelis–Menten + offset | 8 | 4.6 pp | 0 |
| [HLE w/o tools → AA-HLE](https://benchgap.net/calibration/hlenotools/current/to/aa-hle/current.md) | offset logistic | 36 | 4.6 pp | 6 |
| [Agents' Last Exam → AA Tau3 Banking](https://benchgap.net/calibration/agentslastexam/current/to/aa-tau3-banking/current.md) | Hill | 6 | 4.6 pp | 1 |
| [CWE-bench → AA-AnalystAgent](https://benchgap.net/calibration/cwe-bench-v2/current/to/aa-analyst-agent/current.md) | Michaelis–Menten | 9 | 4.6 pp | 6 |
| [Decision Index Knowledge → Artificial Analysis Decision Index v1.0](https://benchgap.net/calibration/jevdecisionindex03knowledge/current/to/aa-decision-index/current.md) | Michaelis–Menten | 11 | 4.7 pp | 12 |
| [Decision Index Language → Decision Index Knowledge](https://benchgap.net/calibration/jevdecisionindex03language/current/to/jevdecisionindex03knowledge/current.md) | inverse Michaelis–Menten | 110 | 4.7 pp | 0 |
| [ERQA → Video-MME (with subtitle)](https://benchgap.net/calibration/erqa/current/to/videommewithsub/current.md) | Michaelis–Menten | 8 | 4.7 pp | 0 |
| [Artificial Analysis Intelligence Index → CritPt](https://benchgap.net/calibration/artificialanalysis/current/to/critpt/current.md) | offset logistic | 154 | 4.7 pp | 6 |
| [StealthMark → FACTS Multimodal](https://benchgap.net/calibration/stealthmark/current/to/facts-multimodal/current.md) | Hill | 17 | 4.7 pp | 7 |
| [Vals SWE-bench → NL2Repo](https://benchgap.net/calibration/valsswebench/current/to/nl2repo/current.md) | Hill | 14 | 4.7 pp | 55 |
| [Artificial Analysis Decision Index v1.0 → Decision Index Tools](https://benchgap.net/calibration/aa-decision-index/current/to/jevdecisionindex03tools/current.md) | Michaelis–Menten | 11 | 4.7 pp | 8 |
| [MCP Atlas → BFCL v4](https://benchgap.net/calibration/mcpatlas/current/to/bfcl-v4/current.md) | Hill | 7 | 4.7 pp | 17 |
| [CompileBench → Mystery Game Puzzles](https://benchgap.net/calibration/compilebench/current/to/mysterygamepuzzles/current.md) | linear | 5 | 4.7 pp | 9 |
| [Decision Index Language → Decision Index Full Score](https://benchgap.net/calibration/jevdecisionindex03language/current/to/decisionindex/current.md) | linear | 110 | 4.7 pp | 0 |
| [SWE-bench Pro → Vals LiveCodeBench](https://benchgap.net/calibration/swe-bench-pro/current/to/valslivecodebench/current.md) | Hill | 36 | 4.7 pp | 18 |
| [LiveCodeBench v6 → AA-SciCode](https://benchgap.net/calibration/livecodebench-v6/current/to/aa-sci-code/current.md) | inverse Michaelis–Menten | 21 | 4.7 pp | 10 |
| [Artificial Analysis Decision Index → Decision Index Tools](https://benchgap.net/calibration/aa-decision-index-v2/current/to/jevdecisionindex03tools/current.md) | Michaelis–Menten + offset | 11 | 4.7 pp | 5 |
| [Vals LiveCodeBench → SWE-bench Pro](https://benchgap.net/calibration/valslivecodebench/current/to/swe-bench-pro/current.md) | offset logistic | 36 | 4.7 pp | 10 |
| [CursorBench 4.0 → FrontierCode 2.0](https://benchgap.net/calibration/cursor-bench40/current/to/frontiercode-2/current.md) | offset logistic | 13 | 4.7 pp | 8 |
| [When2Call (Decision Index) → Decision Index Tools](https://benchgap.net/calibration/aa-decision-when2-call/current/to/jevdecisionindex03tools/current.md) | Michaelis–Menten | 11 | 4.7 pp | 0 |
| [AA Harvey LAB v1.0 → MazeBench](https://benchgap.net/calibration/aa-harvey-lab/current/to/mazebench/current.md) | offset logistic | 11 | 4.7 pp | 0 |
| [Decision Index Full Score → Artificial Analysis Decision Index v1.0](https://benchgap.net/calibration/decisionindex/current/to/aa-decision-index/current.md) | Michaelis–Menten | 11 | 4.7 pp | 0 |
| [OSWorld-Verified → BrowseComp](https://benchgap.net/calibration/osworld-verified/current/to/browsecomp/current.md) | offset logistic | 15 | 4.7 pp | 8 |
| [AA Coding Index → LiveCodeBench v6](https://benchgap.net/calibration/aa-coding-index/current/to/livecodebench-v6/current.md) | Michaelis–Menten | 21 | 4.7 pp | 32 |
| [C-Eval → GPQA](https://benchgap.net/calibration/ceval/current/to/gpqa/current.md) | linear | 11 | 4.7 pp | 0 |
| [HLE-MultiChoice (Decision Index) → Decision Index Arts](https://benchgap.net/calibration/aa-decision-hle-multi-choice/current/to/jevdecisionindex03arts/current.md) | offset logistic | 11 | 4.7 pp | 0 |
| [StealthMark → FrontierPhysics](https://benchgap.net/calibration/stealthmark/current/to/frontierphysics/current.md) | Hill | 20 | 4.7 pp | 18 |
| [Bug Hunt Bench → CursorBench 4.0](https://benchgap.net/calibration/bug-hunt-bench/current/to/cursor-bench40/current.md) | linear | 17 | 4.7 pp | 2 |
| [CyberGym → OSWorld-Verified](https://benchgap.net/calibration/cybergym/current/to/osworld-verified/current.md) | Michaelis–Menten | 10 | 4.7 pp | 13 |
| [StealthMark → DAYJOB Healthcare](https://benchgap.net/calibration/stealthmark/current/to/dayjob-healthcare/current.md) | linear | 12 | 4.7 pp | 11 |
| [AA Coding Index → NL2Repo](https://benchgap.net/calibration/aa-coding-index/current/to/nl2repo/current.md) | linear | 15 | 4.7 pp | 35 |
| [AA Tau3 Banking → BrowseComp](https://benchgap.net/calibration/aa-tau3-banking/current/to/browsecomp/current.md) | linear | 10 | 4.7 pp | 1 |
| [Vals SWE-bench → React Native Evals](https://benchgap.net/calibration/valsswebench/current/to/reactnativeevals/current.md) | Michaelis–Menten + offset | 7 | 4.7 pp | 6 |
| [FrontierPhysics → DAYJOB Healthcare](https://benchgap.net/calibration/frontierphysics/current/to/dayjob-healthcare/current.md) | Michaelis–Menten + offset | 12 | 4.7 pp | 1 |
| [OSWorld-Verified → AA ITBench](https://benchgap.net/calibration/osworld-verified/current/to/aa-itbench/current.md) | Michaelis–Menten + offset | 5 | 4.7 pp | 7 |
| [MazeBench → GDP.pdf](https://benchgap.net/calibration/mazebench/current/to/aa-gdp-pdf/current.md) | Hill | 18 | 4.7 pp | 2 |
| [AA Coding Index → AA LiveCodeBench](https://benchgap.net/calibration/aa-coding-index/current/to/aa-live-code-bench/current.md) | offset logistic | 5 | 4.7 pp | 36 |
| [PolyMath → MAXIFE](https://benchgap.net/calibration/polymath/current/to/maxife/current.md) | Michaelis–Menten | 5 | 4.7 pp | 1 |
| [ProgramBench → CursorBench 4.0](https://benchgap.net/calibration/programbench/current/to/cursor-bench40/current.md) | Michaelis–Menten | 8 | 4.7 pp | 6 |
| [AA-SciCode → React Native Evals](https://benchgap.net/calibration/aa-sci-code/current/to/reactnativeevals/current.md) | linear | 6 | 4.7 pp | 29 |
| [HLE w/ tools → WideResearch](https://benchgap.net/calibration/hlewithtools/current/to/wideresearch/current.md) | linear | 7 | 4.8 pp | 7 |
| [Terminal-Bench 3.0 → Terminal-Bench 4.1](https://benchgap.net/calibration/terminal-bench%23terminal-bench-3-0/current/to/terminal-bench41/current.md) | linear | 10 | 4.8 pp | 3 |
| [FrontierSWE v2 → FrontierCode 2.0](https://benchgap.net/calibration/frontierswev2/current/to/frontiercode-2/current.md) | Hill | 15 | 4.8 pp | 6 |
| [RealWorldQA → ScreenSpot Pro](https://benchgap.net/calibration/realworldqa/current/to/screenspot-pro/current.md) | linear | 7 | 4.8 pp | 9 |
| [NYT Connections → Cookie-Bench](https://benchgap.net/calibration/nyt-connections/current/to/cookie-bench/current.md) | Hill | 5 | 4.8 pp | 10 |
| [ExploitBench Internal Port → Furniture Assembly](https://benchgap.net/calibration/exploitbenchinternalport/current/to/furnitureassembly/current.md) | offset logistic | 7 | 4.8 pp | 2 |
| [AA MMLU-Pro → AA-Omniscience Accuracy](https://benchgap.net/calibration/aa-mmlu-pro/current/to/omniscienceaccuracy/current.md) | linear | 11 | 4.8 pp | 2 |
| [ExploitGym → AA AutomationBench](https://benchgap.net/calibration/exploitgym/current/to/aa-automation-bench/current.md) | Michaelis–Menten | 7 | 4.8 pp | 16 |
| [AA Agentic Index → Hermes Bench](https://benchgap.net/calibration/aa-agentic-index/current/to/hermesbench/current.md) | linear | 5 | 4.8 pp | 0 |
| [Bug Hunt Bench → PostTrainBench v1.1](https://benchgap.net/calibration/bug-hunt-bench/current/to/posttrainbench%23posttrainbench-v1-1/current.md) | Michaelis–Menten | 7 | 4.8 pp | 1 |
| [SWE-Rebench → SWE-bench Verified](https://benchgap.net/calibration/swe-rebench/current/to/swe-bench-verified/current.md) | Michaelis–Menten | 8 | 4.8 pp | 2 |
| [DeepSWE → SWE-Atlas Codebase QnA](https://benchgap.net/calibration/deepswe/current/to/swe-atlas-codebase-qna/current.md) | Hill | 8 | 4.8 pp | 58 |
| [APEX-Agents → BrowseComp](https://benchgap.net/calibration/apexagents/current/to/browsecomp/current.md) | offset logistic | 6 | 4.8 pp | 3 |
| [Vals MMLU-Pro → MMLU-Pro](https://benchgap.net/calibration/vals-mmlu-pro/current/to/mmlu-pro/current.md) | Michaelis–Menten | 9 | 4.8 pp | 37 |
| [FrontierCode 2.0 → PostTrainBench v1.1](https://benchgap.net/calibration/frontiercode-2/current/to/posttrainbench%23posttrainbench-v1-1/current.md) | linear | 9 | 4.8 pp | 2 |
| [PostTrainBench v1.1 → ProgramBench](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1-1/current/to/programbench/current.md) | offset logistic | 5 | 4.8 pp | 10 |
| [AA Coding Index → CursorBench 3.1](https://benchgap.net/calibration/aa-coding-index/current/to/cursor-bench31/current.md) | offset logistic | 5 | 4.8 pp | 9 |
| [FrontierSWE v2 → SWE-bench Pro](https://benchgap.net/calibration/frontierswev2/current/to/swe-bench-pro/current.md) | linear | 9 | 4.8 pp | 2 |
| [Hermes Index → MazeBench](https://benchgap.net/calibration/hermesindex/current/to/mazebench/current.md) | linear | 13 | 4.8 pp | 1 |
| [FrontierSWE v2 → Vals SWE-bench](https://benchgap.net/calibration/frontierswev2/current/to/valsswebench/current.md) | Michaelis–Menten + offset | 12 | 4.9 pp | 1 |
| [CyberGym → HLE w/ tools](https://benchgap.net/calibration/cybergym/current/to/hlewithtools/current.md) | offset logistic | 7 | 4.9 pp | 17 |
| [SWE-bench Pro → OpenHarmony Bench v1.0](https://benchgap.net/calibration/swe-bench-pro/current/to/openharmonybench%23openharmonybench-v1-0/current.md) | offset logistic | 7 | 4.9 pp | 14 |
| [AA-SciCode → CursorBench 4.0](https://benchgap.net/calibration/aa-sci-code/current/to/cursor-bench40/current.md) | inverse Michaelis–Menten | 17 | 4.9 pp | 24 |
| [HealthBench (raw) → ARC-AGI-1](https://benchgap.net/calibration/healthbench/current/to/arcagi1/current.md) | linear | 6 | 4.9 pp | 0 |
| [OSWorld 2.0 → JobBench](https://benchgap.net/calibration/osworld2/current/to/jobbench/current.md) | Michaelis–Menten | 7 | 4.9 pp | 17 |
| [MathVision → Video-MME (w/o subtitle)](https://benchgap.net/calibration/mathvision/current/to/videommenosub/current.md) | Michaelis–Menten + offset | 6 | 4.9 pp | 3 |
| [LiveCodeBench v6 → SciCode](https://benchgap.net/calibration/livecodebench-v6/current/to/scicode/current.md) | inverse Michaelis–Menten | 14 | 4.9 pp | 11 |
| [MMMU-Pro → PerceptionBench](https://benchgap.net/calibration/mmmu-pro/current/to/perceptionbench/current.md) | Michaelis–Menten | 6 | 4.9 pp | 27 |
| [Vals MMLU-Pro → HLE w/o tools](https://benchgap.net/calibration/vals-mmlu-pro/current/to/hlenotools/current.md) | offset logistic | 21 | 4.9 pp | 54 |
| [Video-MME (with subtitle) → CharXiv](https://benchgap.net/calibration/videommewithsub/current/to/charxiv/current.md) | Hill | 10 | 4.9 pp | 1 |
| [WideResearch → SkillsBench](https://benchgap.net/calibration/wideresearch/current/to/skills-bench/current.md) | Hill | 6 | 4.9 pp | 2 |
| [AA AutomationBench → MKA](https://benchgap.net/calibration/aa-automation-bench/current/to/mka/current.md) | Hill | 16 | 4.9 pp | 6 |
| [APEX-Agents-AA → APEX-Agents](https://benchgap.net/calibration/apexagentsaa/current/to/apexagents/current.md) | Michaelis–Menten + offset | 5 | 4.9 pp | 18 |
| [OSWorld 2.1 → HLE w/ tools](https://benchgap.net/calibration/osworld-2-1/current/to/hlewithtools/current.md) | Hill | 8 | 4.9 pp | 4 |
| [CharXiv → AA-MMMU-Pro](https://benchgap.net/calibration/charxiv/current/to/aa-mmmu-pro/current.md) | linear | 31 | 4.9 pp | 5 |
| [AA-SciCode → LiveCodeBench v6](https://benchgap.net/calibration/aa-sci-code/current/to/livecodebench-v6/current.md) | linear | 21 | 4.9 pp | 22 |
| [Artificial Analysis Intelligence Index → HLE w/o tools](https://benchgap.net/calibration/artificialanalysis/current/to/hlenotools/current.md) | linear | 34 | 4.9 pp | 83 |
| [HLE-MultiChoice (Decision Index) → Decision Index Tools](https://benchgap.net/calibration/aa-decision-hle-multi-choice/current/to/jevdecisionindex03tools/current.md) | Hill | 11 | 4.9 pp | 0 |
| [LiveCodeBench Pro → SWE-bench Pro](https://benchgap.net/calibration/livecodebench-pro/current/to/swe-bench-pro/current.md) | offset logistic | 6 | 4.9 pp | 0 |
| [AA-GPQA Diamond → HealthBench Professional](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/healthbenchprofessional/current.md) | offset logistic | 7 | 4.9 pp | 0 |
| [CyberGym → AA ITBench](https://benchgap.net/calibration/cybergym/current/to/aa-itbench/current.md) | Michaelis–Menten | 9 | 4.9 pp | 5 |
| [HealthBench Professional → HLE w/o tools](https://benchgap.net/calibration/healthbenchprofessional/current/to/hlenotools/current.md) | Michaelis–Menten + offset | 6 | 4.9 pp | 2 |
| [SWE-bench Pro → FrontierCode 1.1 Main](https://benchgap.net/calibration/swe-bench-pro/current/to/frontiercode/current.md) | Michaelis–Menten | 9 | 5.0 pp | 12 |
| [ARC-AGI-3 Provider Adapter → CritPt](https://benchgap.net/calibration/arc-agi-3-provider-adapter/current/to/critpt/current.md) | linear | 9 | 5.0 pp | 0 |
| [Agents' Last Exam → BrowseComp](https://benchgap.net/calibration/agentslastexam/current/to/browsecomp/current.md) | Hill | 6 | 5.0 pp | 1 |
| [CursorBench 3.1 → CursorBench 3.2](https://benchgap.net/calibration/cursor-bench31/current/to/cursor-bench32/current.md) | Michaelis–Menten | 5 | 5.0 pp | 1 |
| [Bug Hunt Bench → PostTrainBench](https://benchgap.net/calibration/bug-hunt-bench/current/to/post-train-bench12/current.md) | Hill | 13 | 5.0 pp | 1 |
| [MMMU-Pro → ParseBench-2](https://benchgap.net/calibration/mmmu-pro/current/to/parsebench/current.md) | inverse Michaelis–Menten | 5 | 5.0 pp | 44 |
| [DeepSWE → VulcanBench Frontier v4](https://benchgap.net/calibration/deepswe/current/to/vulcanbench/current.md) | Michaelis–Menten + offset | 15 | 5.0 pp | 21 |
| [CyberGym → GDPval-AA](https://benchgap.net/calibration/cybergym/current/to/gdpvalaanormalized/current.md) | Hill | 18 | 5.0 pp | 10 |
| [React Native Evals → SWE-bench Verified](https://benchgap.net/calibration/reactnativeevals/current/to/swe-bench-verified/current.md) | offset logistic | 7 | 5.0 pp | 2 |
| [SimpleVQA → MLVU (M-Avg)](https://benchgap.net/calibration/simplevqa/current/to/mlvuavg/current.md) | Michaelis–Menten | 10 | 5.0 pp | 2 |
| [ARC-AGI-3 → HLE w/o tools](https://benchgap.net/calibration/arcagi3/current/to/hlenotools/current.md) | linear | 11 | 5.0 pp | 0 |
| [APEX-Agents-AA → ApprenticeBench](https://benchgap.net/calibration/apexagentsaa/current/to/apprenticebench/current.md) | linear | 5 | 5.0 pp | 7 |
| [Mystery Game Puzzles → FrontierPhysics](https://benchgap.net/calibration/mysterygamepuzzles/current/to/frontierphysics/current.md) | Hill | 25 | 5.0 pp | 26 |
| [MedXpertQA (Text) → HLE w/o tools](https://benchgap.net/calibration/medxpertqatext/current/to/hlenotools/current.md) | Michaelis–Menten | 5 | 5.0 pp | 0 |
| [Mystery Game Puzzles → FailureSensorIQ](https://benchgap.net/calibration/mysterygamepuzzles/current/to/failuresensoriq/current.md) | Michaelis–Menten + offset | 9 | 5.0 pp | 5 |
| [AA-Omniscience Accuracy → MMLU-Pro](https://benchgap.net/calibration/omniscienceaccuracy/current/to/mmlu-pro/current.md) | Hill | 31 | 5.0 pp | 54 |
| [VulcanBench Frontier v4 → CursorBench 4.0](https://benchgap.net/calibration/vulcanbench/current/to/cursor-bench40/current.md) | Hill | 10 | 5.0 pp | 0 |
| [BFCL v4 → PinchBench](https://benchgap.net/calibration/bfcl-v4/current/to/pinchbench/current.md) | inverse Michaelis–Menten | 5 | 5.0 pp | 15 |
| [FrontierCode 2.0 → ScarfBench](https://benchgap.net/calibration/frontiercode-2/current/to/scarfbench/current.md) | offset logistic | 7 | 5.0 pp | 6 |
| [HLE-MultiChoice (Decision Index) → Decision Index Knowledge](https://benchgap.net/calibration/aa-decision-hle-multi-choice/current/to/jevdecisionindex03knowledge/current.md) | offset logistic | 11 | 5.1 pp | 2 |
| [AA ITBench → Toolathlon](https://benchgap.net/calibration/aa-itbench/current/to/toolathlon/current.md) | Michaelis–Menten | 7 | 5.1 pp | 10 |
| [SuperGPQA → HLE](https://benchgap.net/calibration/supergpqa/current/to/hle/current.md) | offset logistic | 13 | 5.1 pp | 8 |
| [GDPval-AA → AA ITBench](https://benchgap.net/calibration/gdpvalaanormalized/current/to/aa-itbench/current.md) | Hill | 22 | 5.1 pp | 40 |
| [OfficeQA Pro → CharXiv](https://benchgap.net/calibration/officeqapro/current/to/charxiv/current.md) | offset logistic | 6 | 5.1 pp | 9 |
| [WildGuardTest (Decision Index) → Artificial Analysis Decision Index v1.0](https://benchgap.net/calibration/aa-decision-wild-guard-test/current/to/aa-decision-index/current.md) | offset logistic | 19 | 5.1 pp | 0 |
| [GPQA → HealthBench Hard](https://benchgap.net/calibration/gpqa/current/to/healthbench-hard/current.md) | offset logistic | 6 | 5.1 pp | 38 |
| [Vals MMLU-Pro → GPQA](https://benchgap.net/calibration/vals-mmlu-pro/current/to/gpqa/current.md) | linear | 40 | 5.1 pp | 1 |
| [AI2D_TEST → We-Math](https://benchgap.net/calibration/ai2dtest/current/to/wemath/current.md) | Hill | 5 | 5.1 pp | 2 |
| [MLVU (M-Avg) → MMMU](https://benchgap.net/calibration/mlvuavg/current/to/mmmu/current.md) | inverse Michaelis–Menten | 8 | 5.1 pp | 0 |
| [OSWorld 2.1 Offline → CWE-bench](https://benchgap.net/calibration/osworld-2-1-offline/current/to/cwe-bench-v2/current.md) | Michaelis–Menten | 7 | 5.1 pp | 0 |
| [JobBench → MazeBench](https://benchgap.net/calibration/jobbench/current/to/mazebench/current.md) | inverse Michaelis–Menten | 11 | 5.1 pp | 20 |
| [CursorBench 3.2 → VulcanBench v3](https://benchgap.net/calibration/cursor-bench32/current/to/vulcanbench%23vulcanbench-v3/current.md) | Michaelis–Menten + offset | 8 | 5.1 pp | 6 |
| [Decision Index Tools → Decision Index Knowledge](https://benchgap.net/calibration/jevdecisionindex03tools/current/to/jevdecisionindex03knowledge/current.md) | inverse Michaelis–Menten | 110 | 5.1 pp | 0 |
| [ResearchClawBench → OSWorld-Verified](https://benchgap.net/calibration/researchclawbench/current/to/osworld-verified/current.md) | Hill | 10 | 5.1 pp | 3 |
| [HealthBench (raw) → AA-HLE](https://benchgap.net/calibration/healthbench/current/to/aa-hle/current.md) | Michaelis–Menten | 8 | 5.1 pp | 0 |
| [ApprenticeBench → MazeBench](https://benchgap.net/calibration/apprenticebench/current/to/mazebench/current.md) | Michaelis–Menten + offset | 16 | 5.1 pp | 0 |
| [Vals LiveCodeBench → AA-SciCode](https://benchgap.net/calibration/valslivecodebench/current/to/aa-sci-code/current.md) | Michaelis–Menten + offset | 38 | 5.1 pp | 5 |
| [Terminal-Bench-Science 0.1 → Terminal-Bench 4.1](https://benchgap.net/calibration/terminal-bench-science/current/to/terminal-bench41/current.md) | Michaelis–Menten | 12 | 5.1 pp | 0 |
| [SimpleVQA → MMMU-Pro](https://benchgap.net/calibration/simplevqa/current/to/mmmu-pro/current.md) | Hill | 16 | 5.1 pp | 0 |
| [ARC-AGI-2 → CritPt](https://benchgap.net/calibration/arc-agi-2/current/to/critpt/current.md) | Hill | 39 | 5.1 pp | 2 |
| [CWE-bench v1 → GDPval-AA](https://benchgap.net/calibration/cwe-bench%23cwe-bench-v1/current/to/gdpvalaanormalized/current.md) | Hill | 10 | 5.1 pp | 0 |
| [CursorBench 3.2 → PostTrainBench v1.1](https://benchgap.net/calibration/cursor-bench32/current/to/posttrainbench%23posttrainbench-v1-1/current.md) | Hill | 8 | 5.1 pp | 0 |
| [WideResearch → JobBench](https://benchgap.net/calibration/wideresearch/current/to/jobbench/current.md) | linear | 5 | 5.1 pp | 8 |
| [HealthBench Professional (length-adjusted, Artificial Analysis) → CritPt](https://benchgap.net/calibration/aa-health-bench-professional/current/to/critpt/current.md) | Michaelis–Menten | 6 | 5.1 pp | 0 |
| [ARC-AGI-2 → BioBasicsBench](https://benchgap.net/calibration/arc-agi-2/current/to/biobasicsbench/current.md) | offset logistic | 7 | 5.1 pp | 5 |
| [DeepSWE → SWE-bench Verified](https://benchgap.net/calibration/deepswe/current/to/swe-bench-verified/current.md) | Hill | 12 | 5.1 pp | 22 |
| [ScreenSpot Pro → CharXiv w/o tools](https://benchgap.net/calibration/screenspot-pro/current/to/charxivnotools/current.md) | Michaelis–Menten | 5 | 5.1 pp | 4 |
| [DeepsecBench-AA → HLE w/ tools](https://benchgap.net/calibration/aa-deepsec-bench/current/to/hlewithtools/current.md) | offset logistic | 8 | 5.1 pp | 4 |
| [MazeBench → CWE-bench v1](https://benchgap.net/calibration/mazebench/current/to/cwe-bench%23cwe-bench-v1/current.md) | offset logistic | 10 | 5.1 pp | 36 |
| [WildGuardTest (Decision Index) → Decision Index Language](https://benchgap.net/calibration/aa-decision-wild-guard-test/current/to/jevdecisionindex03language/current.md) | offset logistic | 11 | 5.2 pp | 8 |
| [NL2Repo → Vals SWE-bench](https://benchgap.net/calibration/nl2repo/current/to/valsswebench/current.md) | linear | 14 | 5.2 pp | 19 |
| [MirrorCode → DAYJOB Healthcare](https://benchgap.net/calibration/mirrorcode/current/to/dayjob-healthcare/current.md) | linear | 16 | 5.2 pp | 0 |
| [AA Cyber Index → CWE-Bench-AA](https://benchgap.net/calibration/aa-cyber-index/current/to/aa-cwe-bench/current.md) | offset logistic | 21 | 5.2 pp | 0 |
| [CyberGym → BrowseComp](https://benchgap.net/calibration/cybergym/current/to/browsecomp/current.md) | Michaelis–Menten + offset | 15 | 5.2 pp | 4 |
| [WideResearch → Toolathlon](https://benchgap.net/calibration/wideresearch/current/to/toolathlon/current.md) | Hill | 7 | 5.2 pp | 2 |
| [OmniDocBench 1.5 → CharXiv](https://benchgap.net/calibration/omnidocbench15/current/to/charxiv/current.md) | offset logistic | 10 | 5.2 pp | 0 |
| [ScreenSpot Pro → ParseBench-1](https://benchgap.net/calibration/screenspot-pro/current/to/parse-bench-llama-index/current.md) | Michaelis–Menten + offset | 6 | 5.2 pp | 12 |
| [SingleCellBench → Protocols (troubleshooting)](https://benchgap.net/calibration/singlecellbench/current/to/protocolstroubleshooting/current.md) | linear | 5 | 5.2 pp | 0 |
| [Hermes Bench → MKA](https://benchgap.net/calibration/hermesbench/current/to/mka/current.md) | offset logistic | 10 | 5.2 pp | 0 |
| [MMLU-Pro → GPQA Diamond](https://benchgap.net/calibration/mmlu-pro/current/to/gpqa-diamond/current.md) | Hill | 25 | 5.2 pp | 0 |
| [CursorBench 3.1 → Vals SWE-bench](https://benchgap.net/calibration/cursor-bench31/current/to/valsswebench/current.md) | linear | 7 | 5.2 pp | 0 |
| [Vals SWE-bench → SWE-bench Pro](https://benchgap.net/calibration/valsswebench/current/to/swe-bench-pro/current.md) | Michaelis–Menten + offset | 38 | 5.2 pp | 0 |
| [BioMysteryBench (human-solvable) → AA-HLE](https://benchgap.net/calibration/biomysterybenchhumansolvable/current/to/aa-hle/current.md) | Hill | 5 | 5.2 pp | 0 |
| [ARC-AGI-1 → ARC-AGI-2](https://benchgap.net/calibration/arcagi1/current/to/arc-agi-2/current.md) | inverse Michaelis–Menten | 41 | 5.2 pp | 1 |
| [Mystery Game Puzzles → ExploitBench Internal Port](https://benchgap.net/calibration/mysterygamepuzzles/current/to/exploitbenchinternalport/current.md) | Hill | 7 | 5.2 pp | 12 |
| [AI2D_TEST → RealWorldQA](https://benchgap.net/calibration/ai2dtest/current/to/realworldqa/current.md) | Hill | 7 | 5.2 pp | 0 |
| [Toolathlon-Verified → DeepsecBench-AA](https://benchgap.net/calibration/toolathlonverified/current/to/aa-deepsec-bench/current.md) | Hill | 8 | 5.2 pp | 10 |
| [Vals SWE-bench → FrontierCode 1.1 Main](https://benchgap.net/calibration/valsswebench/current/to/frontiercode/current.md) | Hill | 9 | 5.2 pp | 7 |
| [MMMU → ERQA](https://benchgap.net/calibration/mmmu/current/to/erqa/current.md) | offset logistic | 6 | 5.2 pp | 16 |
| [CWE-bench → CWE-bench v1](https://benchgap.net/calibration/cwe-bench-v2/current/to/cwe-bench%23cwe-bench-v1/current.md) | Michaelis–Menten | 7 | 5.2 pp | 2 |
| [HLE w/ tools → CWE-Bench-AA](https://benchgap.net/calibration/hlewithtools/current/to/aa-cwe-bench/current.md) | Hill | 8 | 5.2 pp | 23 |
| [DeepsecBench-AA → MKA](https://benchgap.net/calibration/aa-deepsec-bench/current/to/mka/current.md) | linear | 11 | 5.2 pp | 1 |
| [Decision Index Language → Artificial Analysis Decision Index v1.0](https://benchgap.net/calibration/jevdecisionindex03language/current/to/aa-decision-index/current.md) | Michaelis–Menten | 11 | 5.3 pp | 2 |
| [SWE-bench Verified → AA Coding Index](https://benchgap.net/calibration/swe-bench-verified/current/to/aa-coding-index/current.md) | offset logistic | 34 | 5.3 pp | 61 |
| [ERQA → CharXiv](https://benchgap.net/calibration/erqa/current/to/charxiv/current.md) | Michaelis–Menten | 17 | 5.3 pp | 0 |
| [GDPval-AA → HLE w/ tools](https://benchgap.net/calibration/gdpvalaanormalized/current/to/hlewithtools/current.md) | linear | 19 | 5.3 pp | 74 |
| [SciCode → Vals SWE-bench](https://benchgap.net/calibration/scicode/current/to/valsswebench/current.md) | Hill | 8 | 5.3 pp | 21 |
| [Vals GPQA Diamond → CritPt](https://benchgap.net/calibration/vals-gpqa-diamond/current/to/critpt/current.md) | offset logistic | 70 | 5.3 pp | 8 |
| [AA-SciCode → CursorBench 3.2](https://benchgap.net/calibration/aa-sci-code/current/to/cursor-bench32/current.md) | linear | 17 | 5.3 pp | 19 |
| [Decision Index Tools → Decision Index Full Score](https://benchgap.net/calibration/jevdecisionindex03tools/current/to/decisionindex/current.md) | Hill | 110 | 5.3 pp | 0 |
| [GPQA → HLE w/o tools](https://benchgap.net/calibration/gpqa/current/to/hlenotools/current.md) | offset logistic | 30 | 5.3 pp | 31 |
| [WildGuardTest (Decision Index) → Decision Index Tools](https://benchgap.net/calibration/aa-decision-wild-guard-test/current/to/jevdecisionindex03tools/current.md) | Hill | 11 | 5.3 pp | 0 |
| [ExploitBench Internal Port → Chess Puzzles](https://benchgap.net/calibration/exploitbenchinternalport/current/to/chesspuzzles/current.md) | Hill | 6 | 5.3 pp | 2 |
| [MMMU-Pro → DynaMath](https://benchgap.net/calibration/mmmu-pro/current/to/dynamath/current.md) | linear | 7 | 5.3 pp | 11 |
| [VulcanBench Frontier v4 → PostTrainBench v1.1](https://benchgap.net/calibration/vulcanbench/current/to/posttrainbench%23posttrainbench-v1-1/current.md) | linear | 5 | 5.3 pp | 1 |
| [ERQA → MLVU (M-Avg)](https://benchgap.net/calibration/erqa/current/to/mlvuavg/current.md) | linear | 8 | 5.3 pp | 0 |
| [AA Agentic Index → CyberGym](https://benchgap.net/calibration/aa-agentic-index/current/to/cybergym/current.md) | linear | 10 | 5.3 pp | 46 |
| [WideResearch → HLE w/ tools](https://benchgap.net/calibration/wideresearch/current/to/hlewithtools/current.md) | linear | 7 | 5.3 pp | 0 |
| [Harvey LAB-AA v1.1 → AA Cyber Index](https://benchgap.net/calibration/aa-harvey-lab-v11/current/to/aa-cyber-index/current.md) | offset logistic | 6 | 5.3 pp | 4 |
| [ParseBench-2 → MMMU-Pro](https://benchgap.net/calibration/parsebench/current/to/mmmu-pro/current.md) | Michaelis–Menten + offset | 5 | 5.3 pp | 6 |
| [AA-HLE → HLE w/o tools](https://benchgap.net/calibration/aa-hle/current/to/hlenotools/current.md) | linear | 36 | 5.4 pp | 25 |
| [Vals SWE-bench → CursorBench 3.2](https://benchgap.net/calibration/valsswebench/current/to/cursor-bench32/current.md) | Michaelis–Menten | 17 | 5.4 pp | 9 |
| [WildGuardTest (Decision Index) → Decision Index Full Score](https://benchgap.net/calibration/aa-decision-wild-guard-test/current/to/decisionindex/current.md) | inverse Michaelis–Menten | 11 | 5.4 pp | 2 |
| [CursorBench 3.2 → FrontierCode 2.0](https://benchgap.net/calibration/cursor-bench32/current/to/frontiercode-2/current.md) | Hill | 6 | 5.4 pp | 5 |
| [Artificial Analysis Intelligence Index → MMLU-Pro](https://benchgap.net/calibration/artificialanalysis/current/to/mmlu-pro/current.md) | Hill | 29 | 5.4 pp | 9 |
| [DAYJOB Healthcare → StealthMark](https://benchgap.net/calibration/dayjob-healthcare/current/to/stealthmark/current.md) | Michaelis–Menten + offset | 12 | 5.4 pp | 0 |
| [RealWorldQA → OCRBench V2](https://benchgap.net/calibration/realworldqa/current/to/ocrbenchv2/current.md) | inverse Michaelis–Menten | 6 | 5.4 pp | 6 |
| [Terminal-Bench Hard → Terminal-Bench 2.0](https://benchgap.net/calibration/terminal-bench-hard/current/to/terminal-bench%23terminal-bench-2-0/current.md) | Hill | 12 | 5.4 pp | 10 |
| [AA Terminal-Bench 4.0 → Terminal-Bench 2.1](https://benchgap.net/calibration/aa-terminal-bench4/current/to/terminal-bench%23terminal-bench-2-1/current.md) | Hill | 5 | 5.4 pp | 18 |
| [OfficeQA Pro → ParseBench-1](https://benchgap.net/calibration/officeqapro/current/to/parse-bench-llama-index/current.md) | Hill | 9 | 5.4 pp | 3 |
| [GPQA Diamond → MMLU-Pro](https://benchgap.net/calibration/gpqa-diamond/current/to/mmlu-pro/current.md) | Michaelis–Menten | 25 | 5.4 pp | 5 |
| [SEC-Bench Pro → SpireClimb](https://benchgap.net/calibration/secbenchpro/current/to/spireclimb/current.md) | Hill | 6 | 5.4 pp | 6 |
| [WideResearch → BrowseComp](https://benchgap.net/calibration/wideresearch/current/to/browsecomp/current.md) | offset logistic | 11 | 5.4 pp | 0 |
| [MMLU-Redux → AA-Omniscience Accuracy](https://benchgap.net/calibration/mmluredux/current/to/omniscienceaccuracy/current.md) | linear | 5 | 5.4 pp | 4 |
| [Hermes Index → GDP.pdf](https://benchgap.net/calibration/hermesindex/current/to/aa-gdp-pdf/current.md) | linear | 6 | 5.4 pp | 0 |
| [ParseBench-2 → AA-MMMU-Pro](https://benchgap.net/calibration/parsebench/current/to/aa-mmmu-pro/current.md) | Michaelis–Menten | 14 | 5.4 pp | 4 |
| [Artificial Analysis Decision Index → When2Call (Decision Index)](https://benchgap.net/calibration/aa-decision-index-v2/current/to/aa-decision-when2-call/current.md) | Hill | 14 | 5.5 pp | 5 |
| [Toolathlon-Verified → OSWorld 2.0](https://benchgap.net/calibration/toolathlonverified/current/to/osworld2/current.md) | offset logistic | 5 | 5.5 pp | 11 |
| [CursorBench 3.2 → Vals SWE-bench](https://benchgap.net/calibration/cursor-bench32/current/to/valsswebench/current.md) | Michaelis–Menten | 17 | 5.5 pp | 0 |
| [RealWorldQA → We-Math](https://benchgap.net/calibration/realworldqa/current/to/wemath/current.md) | Hill | 6 | 5.5 pp | 3 |
| [EBR-bench → DAYJOB Healthcare](https://benchgap.net/calibration/ebrbench/current/to/dayjob-healthcare/current.md) | linear | 11 | 5.5 pp | 0 |
| [τ²-bench results → HLE w/ tools](https://benchgap.net/calibration/tau2-bench/current/to/hlewithtools/current.md) | Hill | 8 | 5.5 pp | 27 |
| [ExploitGym → Hermes Index](https://benchgap.net/calibration/exploitgym/current/to/hermesindex/current.md) | Michaelis–Menten | 6 | 5.5 pp | 7 |
| [WideResearch → Toolathlon-Verified](https://benchgap.net/calibration/wideresearch/current/to/toolathlonverified/current.md) | inverse Michaelis–Menten | 6 | 5.5 pp | 3 |
| [MMMU-Pro w/ Python → ParseBench-1](https://benchgap.net/calibration/mmmupropython/current/to/parse-bench-llama-index/current.md) | linear | 6 | 5.5 pp | 2 |
| [Gert Labs → OSWorld-Verified](https://benchgap.net/calibration/gertlabs/current/to/osworld-verified/current.md) | Hill | 16 | 5.5 pp | 19 |
| [AA Tau3 Banking → AA Agentic Index](https://benchgap.net/calibration/aa-tau3-banking/current/to/aa-agentic-index/current.md) | inverse Michaelis–Menten | 15 | 5.5 pp | 2 |
| [Harvey LAB-AA v1.1 → AA-AnalystAgent](https://benchgap.net/calibration/aa-harvey-lab-v11/current/to/aa-analyst-agent/current.md) | offset logistic | 6 | 5.5 pp | 2 |
| [Decision Index Full Score → Decision Index Arts](https://benchgap.net/calibration/decisionindex/current/to/jevdecisionindex03arts/current.md) | Michaelis–Menten + offset | 110 | 5.5 pp | 0 |
| [HealthBench (length-adjusted) → HealthBench (raw)](https://benchgap.net/calibration/healthbenchlengthadjusted/current/to/healthbench/current.md) | Hill | 8 | 5.5 pp | 0 |
| [MathVision → ERQA](https://benchgap.net/calibration/mathvision/current/to/erqa/current.md) | Michaelis–Menten | 12 | 5.5 pp | 2 |
| [ExploitGym → AA-AnalystAgent](https://benchgap.net/calibration/exploitgym/current/to/aa-analyst-agent/current.md) | Hill | 10 | 5.5 pp | 8 |
| [MirrorCode → DAYJOB](https://benchgap.net/calibration/mirrorcode/current/to/dayjob/current.md) | linear | 16 | 5.5 pp | 0 |
| [MMMLU → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/mmmlu/current/to/artificialanalysis/current.md) | inverse Michaelis–Menten | 7 | 5.5 pp | 2 |
| [BabaIsBench → DAYJOB](https://benchgap.net/calibration/babaisbench/current/to/dayjob/current.md) | offset logistic | 10 | 5.5 pp | 1 |
| [AA-LCR → LongBench v2](https://benchgap.net/calibration/lcr/current/to/longbench-v2/current.md) | Hill | 17 | 5.5 pp | 194 |
| [DeepsecBench-AA → AA AutomationBench](https://benchgap.net/calibration/aa-deepsec-bench/current/to/aa-automation-bench/current.md) | Hill | 11 | 5.5 pp | 6 |
| [τ³-bench results → PinchBench](https://benchgap.net/calibration/tau3-bench/current/to/pinchbench/current.md) | Michaelis–Menten + offset | 7 | 5.5 pp | 9 |
| [Bug Hunt Bench → VulcanBench Frontier v4](https://benchgap.net/calibration/bug-hunt-bench/current/to/vulcanbench/current.md) | offset logistic | 13 | 5.5 pp | 0 |
| [CWE-bench → OSWorld 2.1 Offline](https://benchgap.net/calibration/cwe-bench-v2/current/to/osworld-2-1-offline/current.md) | Michaelis–Menten | 7 | 5.5 pp | 6 |
| [CWE-bench → GDP.pdf](https://benchgap.net/calibration/cwe-bench-v2/current/to/aa-gdp-pdf/current.md) | Michaelis–Menten | 14 | 5.5 pp | 1 |
| [Decision Index Arts → Decision Index Knowledge](https://benchgap.net/calibration/jevdecisionindex03arts/current/to/jevdecisionindex03knowledge/current.md) | inverse Michaelis–Menten | 110 | 5.5 pp | 0 |
| [CyberGym → Gert Labs](https://benchgap.net/calibration/cybergym/current/to/gertlabs/current.md) | linear | 8 | 5.5 pp | 17 |
| [HLE w/o tools → CritPt](https://benchgap.net/calibration/hlenotools/current/to/critpt/current.md) | linear | 40 | 5.5 pp | 2 |
| [CWE-bench v1 → ExploitGym](https://benchgap.net/calibration/cwe-bench%23cwe-bench-v1/current/to/exploitgym/current.md) | inverse Michaelis–Menten | 6 | 5.5 pp | 2 |
| [PostTrainBench v1.1 → ScarfBench](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1-1/current/to/scarfbench/current.md) | Hill | 6 | 5.5 pp | 3 |
| [MCP Atlas → APEX-Agents-AA](https://benchgap.net/calibration/mcpatlas/current/to/apexagentsaa/current.md) | inverse Michaelis–Menten | 14 | 5.5 pp | 9 |
| [SWE-bench Verified → PostTrainBench v1](https://benchgap.net/calibration/swe-bench-verified/current/to/posttrainbench%23posttrainbench-v1/current.md) | Hill | 6 | 5.5 pp | 19 |
| [AA Terminal-Bench 2.1 → Terminal-Bench 2.1 (Vals AI)](https://benchgap.net/calibration/aa-terminal-bench21/current/to/vals-terminal-bench21/current.md) | inverse Michaelis–Menten | 18 | 5.5 pp | 2 |
| [GDPval-AA → MazeBench](https://benchgap.net/calibration/gdpvalaanormalized/current/to/mazebench/current.md) | offset logistic | 41 | 5.5 pp | 81 |
| [AA Coding Index → Vals SWE-bench](https://benchgap.net/calibration/aa-coding-index/current/to/valsswebench/current.md) | Michaelis–Menten + offset | 49 | 5.5 pp | 37 |
| [BFCL v4 → GDPval-AA](https://benchgap.net/calibration/bfcl-v4/current/to/gdpvalaanormalized/current.md) | linear | 9 | 5.5 pp | 13 |
| [DAYJOB → MaxxBench](https://benchgap.net/calibration/dayjob/current/to/maxxbench/current.md) | Michaelis–Menten | 8 | 5.6 pp | 4 |
| [FrontierCode 1.1 Main → Vibe Code Bench](https://benchgap.net/calibration/frontiercode/current/to/vibecodebench/current.md) | Michaelis–Menten | 5 | 5.6 pp | 12 |
| [Terminal-Bench Hard → Terminal-Bench 2.1 (Vals AI)](https://benchgap.net/calibration/terminal-bench-hard/current/to/vals-terminal-bench21/current.md) | Hill | 13 | 5.6 pp | 9 |
| [DynaMath → ERQA](https://benchgap.net/calibration/dynamath/current/to/erqa/current.md) | Hill | 6 | 5.6 pp | 0 |
| [ParseBench-1 → CharXiv](https://benchgap.net/calibration/parse-bench-llama-index/current/to/charxiv/current.md) | linear | 12 | 5.6 pp | 20 |
| [HLE w/o tools → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/hlenotools/current/to/artificialanalysis/current.md) | linear | 34 | 5.6 pp | 21 |
| [Chess Puzzles → FACTS Multimodal](https://benchgap.net/calibration/chesspuzzles/current/to/facts-multimodal/current.md) | linear | 20 | 5.6 pp | 28 |
| [ARC-AGI-1 → CritPt](https://benchgap.net/calibration/arcagi1/current/to/critpt/current.md) | offset logistic | 37 | 5.6 pp | 0 |
| [Toolathlon-Verified → GDPval-AA](https://benchgap.net/calibration/toolathlonverified/current/to/gdpvalaanormalized/current.md) | inverse Michaelis–Menten | 14 | 5.6 pp | 7 |
| [HMMT Feb 2026 → IMOAnswerBench](https://benchgap.net/calibration/hmmtfeb2026/current/to/imoanswerbench/current.md) | Michaelis–Menten + offset | 9 | 5.6 pp | 18 |
| [BrowseComp → HLE w/ tools](https://benchgap.net/calibration/browsecomp/current/to/hlewithtools/current.md) | Hill | 22 | 5.6 pp | 9 |
| [SWE-bench Pro → CursorBench 3.1](https://benchgap.net/calibration/swe-bench-pro/current/to/cursor-bench31/current.md) | Michaelis–Menten + offset | 5 | 5.6 pp | 36 |
| [Toolathlon → WideResearch](https://benchgap.net/calibration/toolathlon/current/to/wideresearch/current.md) | linear | 7 | 5.6 pp | 1 |
| [SciCode → SWE-bench Verified](https://benchgap.net/calibration/scicode/current/to/swe-bench-verified/current.md) | Hill | 20 | 5.6 pp | 7 |
| [Decision Index Knowledge → Decision Index Language](https://benchgap.net/calibration/jevdecisionindex03knowledge/current/to/jevdecisionindex03language/current.md) | Hill | 110 | 5.6 pp | 0 |
| [Decision Index Retrieval → Decision Index Knowledge](https://benchgap.net/calibration/jevdecisionindex03retrieval/current/to/jevdecisionindex03knowledge/current.md) | inverse Michaelis–Menten | 110 | 5.6 pp | 0 |
| [PostTrainBench v1.1 → CursorBench 3.2](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1-1/current/to/cursor-bench32/current.md) | offset logistic | 8 | 5.6 pp | 0 |
| [Artificial Analysis Decision Index → Decision Index Language](https://benchgap.net/calibration/aa-decision-index-v2/current/to/jevdecisionindex03language/current.md) | inverse Michaelis–Menten | 11 | 5.6 pp | 5 |
| [EBR-bench → FrontierPhysics](https://benchgap.net/calibration/ebrbench/current/to/frontierphysics/current.md) | linear | 15 | 5.6 pp | 0 |
| [CharXiv → ERQA](https://benchgap.net/calibration/charxiv/current/to/erqa/current.md) | Michaelis–Menten | 17 | 5.6 pp | 6 |
| [MMMU-Pro → CharXiv](https://benchgap.net/calibration/mmmu-pro/current/to/charxiv/current.md) | inverse Michaelis–Menten | 30 | 5.6 pp | 11 |
| [Vals LiveCodeBench → VulcanBench v3](https://benchgap.net/calibration/valslivecodebench/current/to/vulcanbench%23vulcanbench-v3/current.md) | offset logistic | 12 | 5.6 pp | 68 |
| [Terminal-Bench 4.1 → Terminal-Bench 2.1](https://benchgap.net/calibration/terminal-bench41/current/to/terminal-bench%23terminal-bench-2-1/current.md) | Michaelis–Menten + offset | 23 | 5.6 pp | 11 |
| [FACTS Multimodal → SEC-Bench Pro](https://benchgap.net/calibration/facts-multimodal/current/to/secbenchpro/current.md) | offset logistic | 8 | 5.6 pp | 13 |
| [ParseBench-1 → SWE-bench Multimodal](https://benchgap.net/calibration/parse-bench-llama-index/current/to/swe-bench-multimodal/current.md) | offset logistic | 5 | 5.6 pp | 31 |
| [Vals SWE-bench → AA Coding Index](https://benchgap.net/calibration/valsswebench/current/to/aa-coding-index/current.md) | Hill | 49 | 5.6 pp | 12 |
| [AA-AnalystAgent → MKA](https://benchgap.net/calibration/aa-analyst-agent/current/to/mka/current.md) | Michaelis–Menten + offset | 12 | 5.6 pp | 6 |
| [BinaryAudit → OTelBench](https://benchgap.net/calibration/binaryaudit/current/to/otelbench/current.md) | Hill | 11 | 5.6 pp | 6 |
| [ResearchClawBench → BrowseComp](https://benchgap.net/calibration/researchclawbench/current/to/browsecomp/current.md) | offset logistic | 10 | 5.6 pp | 4 |
| [APEX-Agents → APEX-Agents-AA](https://benchgap.net/calibration/apexagents/current/to/apexagentsaa/current.md) | Michaelis–Menten | 5 | 5.7 pp | 0 |
| [SimpleVQA → Video-MME (with subtitle)](https://benchgap.net/calibration/simplevqa/current/to/videommewithsub/current.md) | Michaelis–Menten | 10 | 5.7 pp | 0 |
| [Artificial Analysis Decision Index v1.0 → Decision Index Knowledge](https://benchgap.net/calibration/aa-decision-index/current/to/jevdecisionindex03knowledge/current.md) | inverse Michaelis–Menten | 11 | 5.7 pp | 0 |
| [Decision Index Retrieval → Artificial Analysis Decision Index v1.0](https://benchgap.net/calibration/jevdecisionindex03retrieval/current/to/aa-decision-index/current.md) | Michaelis–Menten | 11 | 5.7 pp | 5 |
| [HLE w/ tools → Toolathlon-Verified](https://benchgap.net/calibration/hlewithtools/current/to/toolathlonverified/current.md) | Michaelis–Menten + offset | 14 | 5.7 pp | 9 |
| [BrowseComp → OSWorld-Verified](https://benchgap.net/calibration/browsecomp/current/to/osworld-verified/current.md) | Hill | 15 | 5.7 pp | 28 |
| [ScarfBench → DeepSWE](https://benchgap.net/calibration/scarfbench/current/to/deepswe/current.md) | Michaelis–Menten | 9 | 5.7 pp | 0 |
| [BioMysteryBench (human-solvable) → CritPt](https://benchgap.net/calibration/biomysterybenchhumansolvable/current/to/critpt/current.md) | linear | 5 | 5.7 pp | 0 |
| [StealthMark → MaxxBench](https://benchgap.net/calibration/stealthmark/current/to/maxxbench/current.md) | linear | 9 | 5.7 pp | 23 |
| [ExtractBench Pro → ParseBench-1](https://benchgap.net/calibration/extractbench-pro/current/to/parse-bench-llama-index/current.md) | linear | 7 | 5.7 pp | 1 |
| [LiveCodeBench v6 → AA Coding Index](https://benchgap.net/calibration/livecodebench-v6/current/to/aa-coding-index/current.md) | Hill | 21 | 5.7 pp | 15 |
| [Artificial Analysis Intelligence Index → Vals GPQA Diamond](https://benchgap.net/calibration/artificialanalysis/current/to/vals-gpqa-diamond/current.md) | Hill | 72 | 5.7 pp | 8 |
| [VideoMMMU → DynaMath](https://benchgap.net/calibration/videommmu/current/to/dynamath/current.md) | linear | 7 | 5.7 pp | 0 |
| [AA-SciCode → Vals LiveCodeBench](https://benchgap.net/calibration/aa-sci-code/current/to/valslivecodebench/current.md) | Michaelis–Menten | 38 | 5.7 pp | 42 |
| [AA ITBench → BrowseComp](https://benchgap.net/calibration/aa-itbench/current/to/browsecomp/current.md) | Hill | 10 | 5.7 pp | 1 |
| [MMLU-Pro → GPQA](https://benchgap.net/calibration/mmlu-pro/current/to/gpqa/current.md) | Hill | 62 | 5.7 pp | 2 |
| [CountBench → We-Math](https://benchgap.net/calibration/countbench/current/to/wemath/current.md) | linear | 6 | 5.7 pp | 0 |
| [CursorBench 4.0 → FrontierCode 1.1 Main](https://benchgap.net/calibration/cursor-bench40/current/to/frontiercode/current.md) | Michaelis–Menten | 5 | 5.7 pp | 0 |
| [AA Harvey LAB v1.0 → HLE w/ tools](https://benchgap.net/calibration/aa-harvey-lab/current/to/hlewithtools/current.md) | Hill | 5 | 5.7 pp | 0 |
| [ProgramBench → AA-SciCode](https://benchgap.net/calibration/programbench/current/to/aa-sci-code/current.md) | offset logistic | 15 | 5.7 pp | 0 |
| [BioBasicsBench → CritPt](https://benchgap.net/calibration/biobasicsbench/current/to/critpt/current.md) | Hill | 7 | 5.7 pp | 0 |
| [OpenHarmony Bench v1.0 → NL2Repo](https://benchgap.net/calibration/openharmonybench%23openharmonybench-v1-0/current/to/nl2repo/current.md) | linear | 10 | 5.7 pp | 0 |
| [MazeBench → MKA](https://benchgap.net/calibration/mazebench/current/to/mka/current.md) | linear | 18 | 5.7 pp | 15 |
| [AA Coding Index → PostTrainBench v1.1](https://benchgap.net/calibration/aa-coding-index/current/to/posttrainbench%23posttrainbench-v1-1/current.md) | Hill | 12 | 5.7 pp | 6 |
| [GPQA → AA-HLE](https://benchgap.net/calibration/gpqa/current/to/aa-hle/current.md) | inverse Michaelis–Menten | 91 | 5.7 pp | 29 |
| [GDP.pdf → MKA](https://benchgap.net/calibration/aa-gdp-pdf/current/to/mka/current.md) | Hill | 15 | 5.7 pp | 0 |
| [MMMU → DynaMath](https://benchgap.net/calibration/mmmu/current/to/dynamath/current.md) | Michaelis–Menten + offset | 7 | 5.8 pp | 11 |
| [Vals SWE-bench → CursorBench 3.1](https://benchgap.net/calibration/valsswebench/current/to/cursor-bench31/current.md) | linear | 7 | 5.8 pp | 14 |
| [Toolathlon-Verified → HLE w/ tools](https://benchgap.net/calibration/toolathlonverified/current/to/hlewithtools/current.md) | linear | 14 | 5.8 pp | 1 |
| [Hermes Index → AA-AnalystAgent](https://benchgap.net/calibration/hermesindex/current/to/aa-analyst-agent/current.md) | Michaelis–Menten | 9 | 5.8 pp | 1 |
| [Toolathlon → ApprenticeBench](https://benchgap.net/calibration/toolathlon/current/to/apprenticebench/current.md) | Hill | 5 | 5.8 pp | 0 |
| [AA ITBench → CyberGym](https://benchgap.net/calibration/aa-itbench/current/to/cybergym/current.md) | Michaelis–Menten | 9 | 5.8 pp | 2 |
| [CyberGym → AA Agentic Index](https://benchgap.net/calibration/cybergym/current/to/aa-agentic-index/current.md) | inverse Michaelis–Menten | 10 | 5.8 pp | 10 |
| [Vals GPQA Diamond → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/vals-gpqa-diamond/current/to/artificialanalysis/current.md) | inverse Michaelis–Menten | 72 | 5.8 pp | 10 |
| [CursorBench 4.0 → FrontierSWE v2](https://benchgap.net/calibration/cursor-bench40/current/to/frontierswev2/current.md) | linear | 15 | 5.8 pp | 0 |
| [DAYJOB → SEC-Bench Pro](https://benchgap.net/calibration/dayjob/current/to/secbenchpro/current.md) | Michaelis–Menten | 8 | 5.8 pp | 3 |
| [CritPt → AA-HLE](https://benchgap.net/calibration/critpt/current/to/aa-hle/current.md) | Hill | 200 | 5.8 pp | 2 |
| [ApprenticeBench → MCP Atlas](https://benchgap.net/calibration/apprenticebench/current/to/mcpatlas/current.md) | linear | 5 | 5.8 pp | 11 |
| [CWE-bench v1 → CWE-bench](https://benchgap.net/calibration/cwe-bench%23cwe-bench-v1/current/to/cwe-bench-v2/current.md) | Michaelis–Menten | 7 | 5.8 pp | 3 |
| [CountBench → V*](https://benchgap.net/calibration/countbench/current/to/vstar/current.md) | inverse Michaelis–Menten | 6 | 5.8 pp | 3 |
| [Terminal-Bench 2.1 (Vals AI) → Terminal-Bench 2.0](https://benchgap.net/calibration/vals-terminal-bench21/current/to/terminal-bench%23terminal-bench-2-0/current.md) | Michaelis–Menten | 19 | 5.8 pp | 43 |
| [FrontierCode 2.0 → VulcanBench Frontier v4](https://benchgap.net/calibration/frontiercode-2/current/to/vulcanbench/current.md) | linear | 14 | 5.8 pp | 2 |
| [HLE-Verified → AA-Omniscience Accuracy](https://benchgap.net/calibration/hleverified/current/to/omniscienceaccuracy/current.md) | Michaelis–Menten + offset | 6 | 5.8 pp | 0 |
| [MLVU (M-Avg) → MMMU-Pro](https://benchgap.net/calibration/mlvuavg/current/to/mmmu-pro/current.md) | Hill | 10 | 5.8 pp | 0 |
| [AA Terminal-Bench 4.0 → Terminal-Bench 4.1](https://benchgap.net/calibration/aa-terminal-bench4/current/to/terminal-bench41/current.md) | linear | 19 | 5.8 pp | 3 |
| [Mystery Game Puzzles → FACTS Multimodal](https://benchgap.net/calibration/mysterygamepuzzles/current/to/facts-multimodal/current.md) | linear | 21 | 5.8 pp | 3 |
| [Gert Labs → WideResearch](https://benchgap.net/calibration/gertlabs/current/to/wideresearch/current.md) | Hill | 7 | 5.9 pp | 22 |
| [AA Coding Index → PostTrainBench](https://benchgap.net/calibration/aa-coding-index/current/to/post-train-bench12/current.md) | Hill | 12 | 5.9 pp | 4 |
| [MMVU → CharXiv](https://benchgap.net/calibration/mmvu/current/to/charxiv/current.md) | inverse Michaelis–Menten | 6 | 5.9 pp | 0 |
| [VideoMMMU → CharXiv](https://benchgap.net/calibration/videommmu/current/to/charxiv/current.md) | Hill | 10 | 5.9 pp | 0 |
| [Gert Labs → OSWorld 2.0](https://benchgap.net/calibration/gertlabs/current/to/osworld2/current.md) | linear | 5 | 5.9 pp | 22 |
| [Decision Index Full Score → Decision Index Tools](https://benchgap.net/calibration/decisionindex/current/to/jevdecisionindex03tools/current.md) | Hill | 110 | 5.9 pp | 0 |
| [Agents' Last Exam → CyberGym](https://benchgap.net/calibration/agentslastexam/current/to/cybergym/current.md) | Michaelis–Menten | 9 | 5.9 pp | 1 |
| [Vals MMLU-Pro → CritPt](https://benchgap.net/calibration/vals-mmlu-pro/current/to/critpt/current.md) | inverse Michaelis–Menten | 70 | 5.9 pp | 0 |
| [MazeBench → AA-AnalystAgent](https://benchgap.net/calibration/mazebench/current/to/aa-analyst-agent/current.md) | Hill | 22 | 5.9 pp | 13 |
| [SWE-bench Pro → SciCode](https://benchgap.net/calibration/swe-bench-pro/current/to/scicode/current.md) | Michaelis–Menten | 20 | 5.9 pp | 3 |
| [τ³-bench results → AA Agentic Index](https://benchgap.net/calibration/tau3-bench/current/to/aa-agentic-index/current.md) | offset logistic | 8 | 5.9 pp | 10 |
| [DAYJOB Healthcare → FrontierPhysics](https://benchgap.net/calibration/dayjob-healthcare/current/to/frontierphysics/current.md) | linear | 12 | 5.9 pp | 0 |
| [HealthBench (length-adjusted) → AA-HLE](https://benchgap.net/calibration/healthbenchlengthadjusted/current/to/aa-hle/current.md) | Michaelis–Menten + offset | 8 | 5.9 pp | 0 |
| [FrontierCode 1.1 Main → PostTrainBench v1.1](https://benchgap.net/calibration/frontiercode/current/to/posttrainbench%23posttrainbench-v1-1/current.md) | Hill | 6 | 5.9 pp | 0 |
| [Chartography (tools) → Chartography (no tools)](https://benchgap.net/calibration/chartographywithtools/current/to/chartography/current.md) | inverse Michaelis–Menten | 7 | 6.0 pp | 3 |
| [AutomationBench → CWE-Bench-AA](https://benchgap.net/calibration/automationbench/current/to/aa-cwe-bench/current.md) | Hill | 10 | 6.0 pp | 7 |
| [ARC-AGI-1 → AA-HLE](https://benchgap.net/calibration/arcagi1/current/to/aa-hle/current.md) | offset logistic | 36 | 6.0 pp | 2 |
| [DeepSWE → NL2Repo](https://benchgap.net/calibration/deepswe/current/to/nl2repo/current.md) | linear | 15 | 6.0 pp | 36 |
| [FrontierPhysics → StealthMark](https://benchgap.net/calibration/frontierphysics/current/to/stealthmark/current.md) | linear | 20 | 6.0 pp | 6 |
| [CWE-Bench-AA → HLE w/ tools](https://benchgap.net/calibration/aa-cwe-bench/current/to/hlewithtools/current.md) | Michaelis–Menten | 8 | 6.0 pp | 0 |
| [LiveCodeBench v6 → SWE-bench Pro](https://benchgap.net/calibration/livecodebench-v6/current/to/swe-bench-pro/current.md) | linear | 22 | 6.0 pp | 15 |
| [GDPval-AA → QwenClawBench](https://benchgap.net/calibration/gdpvalaanormalized/current/to/qwenclawbench/current.md) | offset logistic | 6 | 6.0 pp | 76 |
| [JobBench → ApprenticeBench](https://benchgap.net/calibration/jobbench/current/to/apprenticebench/current.md) | Hill | 7 | 6.0 pp | 19 |
| [Toolathlon-Verified → CyberGym](https://benchgap.net/calibration/toolathlonverified/current/to/cybergym/current.md) | Hill | 7 | 6.0 pp | 3 |
| [CursorBench 3.2 → ProgramBench](https://benchgap.net/calibration/cursor-bench32/current/to/programbench/current.md) | Hill | 5 | 6.0 pp | 9 |
| [Decision Index Knowledge → Decision Index Arts](https://benchgap.net/calibration/jevdecisionindex03knowledge/current/to/jevdecisionindex03arts/current.md) | Michaelis–Menten + offset | 110 | 6.0 pp | 0 |
| [FrontierCode 2.0 → cursorBench41](https://benchgap.net/calibration/frontiercode-2/current/to/cursor-bench41/current.md) | Michaelis–Menten | 10 | 6.0 pp | 8 |
| [AA-HLE → HealthBench (raw)](https://benchgap.net/calibration/aa-hle/current/to/healthbench/current.md) | Michaelis–Menten | 8 | 6.0 pp | 109 |
| [HLE → HLE w/o tools](https://benchgap.net/calibration/hle/current/to/hlenotools/current.md) | linear | 36 | 6.0 pp | 4 |
| [GDPval-AA → OSWorld 2.1](https://benchgap.net/calibration/gdpvalaanormalized/current/to/osworld-2-1/current.md) | Michaelis–Menten | 9 | 6.0 pp | 27 |
| [PostTrainBench v1.1 → SWE-bench Pro](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1-1/current/to/swe-bench-pro/current.md) | linear | 10 | 6.0 pp | 0 |
| [SWE-bench Pro → LiveCodeBench Pro](https://benchgap.net/calibration/swe-bench-pro/current/to/livecodebench-pro/current.md) | Michaelis–Menten | 6 | 6.0 pp | 74 |
| [WideResearch → MCP-Tasks](https://benchgap.net/calibration/wideresearch/current/to/mcptasks/current.md) | Hill | 5 | 6.0 pp | 5 |
| [GDPval-AA → CyberGym](https://benchgap.net/calibration/gdpvalaanormalized/current/to/cybergym/current.md) | Michaelis–Menten | 18 | 6.0 pp | 46 |
| [Toolathlon-Verified → BrowseComp](https://benchgap.net/calibration/toolathlonverified/current/to/browsecomp/current.md) | Hill | 11 | 6.0 pp | 2 |
| [PostTrainBench → FrontierCode 2.0](https://benchgap.net/calibration/post-train-bench12/current/to/frontiercode-2/current.md) | Hill | 15 | 6.0 pp | 4 |
| [MedXpertQA (MM) → ERQA](https://benchgap.net/calibration/medxpertqamm/current/to/erqa/current.md) | inverse Michaelis–Menten | 7 | 6.0 pp | 0 |
| [ARC-AGI-2 → AA-HLE](https://benchgap.net/calibration/arc-agi-2/current/to/aa-hle/current.md) | offset logistic | 38 | 6.0 pp | 1 |
| [CyberGym → APEX-Agents-AA](https://benchgap.net/calibration/cybergym/current/to/apexagentsaa/current.md) | inverse Michaelis–Menten | 7 | 6.0 pp | 7 |
| [AA Tau3 Banking → MCP Atlas](https://benchgap.net/calibration/aa-tau3-banking/current/to/mcpatlas/current.md) | Michaelis–Menten + offset | 7 | 6.0 pp | 7 |
| [AA Coding Index → SWE-bench Pro](https://benchgap.net/calibration/aa-coding-index/current/to/swe-bench-pro/current.md) | Michaelis–Menten | 38 | 6.0 pp | 22 |
| [ExploitBench → StealthMark](https://benchgap.net/calibration/exploitbench/current/to/stealthmark/current.md) | Hill | 11 | 6.1 pp | 5 |
| [ProgramBench → Vals LiveCodeBench](https://benchgap.net/calibration/programbench/current/to/valslivecodebench/current.md) | offset logistic | 6 | 6.1 pp | 5 |
| [CursorBench 4.0 → PostTrainBench](https://benchgap.net/calibration/cursor-bench40/current/to/post-train-bench12/current.md) | Hill | 8 | 6.1 pp | 0 |
| [WideResearch → Gert Labs](https://benchgap.net/calibration/wideresearch/current/to/gertlabs/current.md) | Michaelis–Menten | 7 | 6.1 pp | 5 |
| [AA Terminal-Bench 4.0 → Terminal-Bench 4.0](https://benchgap.net/calibration/aa-terminal-bench4/current/to/terminal-bench/current.md) | Hill | 10 | 6.1 pp | 3 |
| [AA-SciCode → AA Coding Index](https://benchgap.net/calibration/aa-sci-code/current/to/aa-coding-index/current.md) | offset logistic | 74 | 6.1 pp | 8 |
| [Toolathlon → AA ITBench](https://benchgap.net/calibration/toolathlon/current/to/aa-itbench/current.md) | inverse Michaelis–Menten | 7 | 6.1 pp | 0 |
| [Chess Puzzles → FrontierPhysics](https://benchgap.net/calibration/chesspuzzles/current/to/frontierphysics/current.md) | linear | 24 | 6.1 pp | 18 |
| [BioBasicsBench → HealthBench Professional](https://benchgap.net/calibration/biobasicsbench/current/to/healthbenchprofessional/current.md) | Michaelis–Menten | 5 | 6.1 pp | 0 |
| [HLE w/ tools → GDP.pdf](https://benchgap.net/calibration/hlewithtools/current/to/aa-gdp-pdf/current.md) | linear | 6 | 6.1 pp | 11 |
| [MazeBench → OSWorld 2.1](https://benchgap.net/calibration/mazebench/current/to/osworld-2-1/current.md) | Michaelis–Menten + offset | 10 | 6.1 pp | 15 |
| [AA-GPQA Diamond → GPQA](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/gpqa/current.md) | linear | 88 | 6.1 pp | 57 |
| [AA AutomationBench → AutomationBench](https://benchgap.net/calibration/aa-automation-bench/current/to/automationbench/current.md) | offset logistic | 6 | 6.1 pp | 16 |
| [NYT Connections → SkillsBench official leaderboard (BenchFlow)](https://benchgap.net/calibration/nyt-connections/current/to/skillsbench/current.md) | Michaelis–Menten | 19 | 6.1 pp | 31 |
| [HLE → CritPt](https://benchgap.net/calibration/hle/current/to/critpt/current.md) | linear | 61 | 6.1 pp | 3 |
| [Chess Puzzles → DAYJOB Healthcare](https://benchgap.net/calibration/chesspuzzles/current/to/dayjob-healthcare/current.md) | Hill | 19 | 6.1 pp | 18 |
| [SWE-bench Pro → NL2Repo](https://benchgap.net/calibration/swe-bench-pro/current/to/nl2repo/current.md) | linear | 28 | 6.1 pp | 8 |
| [ParseBench-1 → AA-MMMU-Pro](https://benchgap.net/calibration/parse-bench-llama-index/current/to/aa-mmmu-pro/current.md) | offset logistic | 29 | 6.1 pp | 3 |
| [ARC-AGI-2 → HLE](https://benchgap.net/calibration/arc-agi-2/current/to/hle/current.md) | Hill | 19 | 6.1 pp | 10 |
| [Agents' Last Exam → MazeBench](https://benchgap.net/calibration/agentslastexam/current/to/mazebench/current.md) | Hill | 12 | 6.2 pp | 1 |
| [SWE-bench Verified → NL2Repo](https://benchgap.net/calibration/swe-bench-verified/current/to/nl2repo/current.md) | Hill | 19 | 6.2 pp | 25 |
| [Terminal-Bench 4.1 → AA Terminal-Bench 4.0](https://benchgap.net/calibration/terminal-bench41/current/to/aa-terminal-bench4/current.md) | linear | 19 | 6.2 pp | 31 |
| [Hermes Bench → CWE-bench](https://benchgap.net/calibration/hermesbench/current/to/cwe-bench-v2/current.md) | linear | 7 | 6.2 pp | 0 |
| [JobBench → BrowseComp](https://benchgap.net/calibration/jobbench/current/to/browsecomp/current.md) | linear | 12 | 6.2 pp | 12 |
| [SEC-Bench Pro → FrontierPhysics](https://benchgap.net/calibration/secbenchpro/current/to/frontierphysics/current.md) | linear | 11 | 6.2 pp | 0 |
| [GPQA → AA-GPQA Diamond](https://benchgap.net/calibration/gpqa/current/to/aa-gpqa-diamond/current.md) | linear | 88 | 6.2 pp | 26 |
| [VideoMMMU → MathVision](https://benchgap.net/calibration/videommmu/current/to/mathvision/current.md) | Michaelis–Menten | 10 | 6.2 pp | 0 |
| [Video-MME (with subtitle) → Video-MME (w/o subtitle)](https://benchgap.net/calibration/videommewithsub/current/to/videommenosub/current.md) | Michaelis–Menten | 7 | 6.2 pp | 0 |
| [Vals LiveCodeBench → FrontierCode 1.1 Main](https://benchgap.net/calibration/valslivecodebench/current/to/frontiercode/current.md) | Hill | 9 | 6.2 pp | 8 |
| [AA-MMMU-Pro → ParseBench-2](https://benchgap.net/calibration/aa-mmmu-pro/current/to/parsebench/current.md) | inverse Michaelis–Menten | 14 | 6.2 pp | 53 |
| [ARC-AGI-3 Provider Adapter → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/arc-agi-3-provider-adapter/current/to/artificialanalysis/current.md) | linear | 13 | 6.2 pp | 0 |
| [JobBench → GDPval-AA](https://benchgap.net/calibration/jobbench/current/to/gdpvalaanormalized/current.md) | linear | 16 | 6.2 pp | 14 |
| [OSWorld 2.0 → MazeBench](https://benchgap.net/calibration/osworld2/current/to/mazebench/current.md) | Michaelis–Menten + offset | 18 | 6.2 pp | 0 |
| [Decision Index Arts → Decision Index Full Score](https://benchgap.net/calibration/jevdecisionindex03arts/current/to/decisionindex/current.md) | linear | 110 | 6.2 pp | 0 |
| [OSWorld-Verified → Claw-Eval](https://benchgap.net/calibration/osworld-verified/current/to/claw-eval/current.md) | Michaelis–Menten | 9 | 6.2 pp | 14 |
| [Chess Puzzles → SkillsBench official leaderboard (BenchFlow)](https://benchgap.net/calibration/chesspuzzles/current/to/skillsbench/current.md) | Michaelis–Menten | 17 | 6.2 pp | 27 |
| [ScarfBench → PostTrainBench v1.1](https://benchgap.net/calibration/scarfbench/current/to/posttrainbench%23posttrainbench-v1-1/current.md) | inverse Michaelis–Menten | 6 | 6.2 pp | 1 |
| [GraphWalks BFS 256K–1M → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/graphwalks-bfs256k1m/current/to/artificialanalysis/current.md) | offset logistic | 5 | 6.2 pp | 0 |
| [AA-MMMU-Pro → ParseBench-1](https://benchgap.net/calibration/aa-mmmu-pro/current/to/parse-bench-llama-index/current.md) | Hill | 29 | 6.2 pp | 60 |
| [EBR-bench → FACTS Multimodal](https://benchgap.net/calibration/ebrbench/current/to/facts-multimodal/current.md) | Hill | 11 | 6.2 pp | 0 |
| [OSWorld 2.1 Offline → MazeBench](https://benchgap.net/calibration/osworld-2-1-offline/current/to/mazebench/current.md) | linear | 11 | 6.2 pp | 0 |
| [Vals GPQA Diamond → AA-HLE](https://benchgap.net/calibration/vals-gpqa-diamond/current/to/aa-hle/current.md) | offset logistic | 74 | 6.2 pp | 8 |
| [AA-MMMU-Pro → CharXiv](https://benchgap.net/calibration/aa-mmmu-pro/current/to/charxiv/current.md) | Michaelis–Menten | 31 | 6.3 pp | 48 |
| [MazeBench → ExploitGym](https://benchgap.net/calibration/mazebench/current/to/exploitgym/current.md) | Michaelis–Menten + offset | 15 | 6.3 pp | 14 |
| [Vals MMLU-Pro → AA-GPQA Diamond](https://benchgap.net/calibration/vals-mmlu-pro/current/to/aa-gpqa-diamond/current.md) | linear | 74 | 6.3 pp | 0 |
| [AA-HLE → Vals GPQA Diamond](https://benchgap.net/calibration/aa-hle/current/to/vals-gpqa-diamond/current.md) | Michaelis–Menten | 74 | 6.3 pp | 0 |
| [APEX-Agents-AA → AA ITBench](https://benchgap.net/calibration/apexagentsaa/current/to/aa-itbench/current.md) | offset logistic | 9 | 6.3 pp | 3 |
| [MMLU-Redux → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/mmluredux/current/to/artificialanalysis/current.md) | Hill | 6 | 6.3 pp | 4 |
| [GDPval-AA → Toolathlon](https://benchgap.net/calibration/gdpvalaanormalized/current/to/toolathlon/current.md) | offset logistic | 19 | 6.3 pp | 76 |
| [MMVU → MathVision](https://benchgap.net/calibration/mmvu/current/to/mathvision/current.md) | inverse Michaelis–Menten | 7 | 6.3 pp | 0 |
| [MCP Atlas → Claw-Eval](https://benchgap.net/calibration/mcpatlas/current/to/claw-eval/current.md) | linear | 19 | 6.3 pp | 20 |
| [Terminal-Bench 4.0 → AA Terminal-Bench 4.0](https://benchgap.net/calibration/terminal-bench/current/to/aa-terminal-bench4/current.md) | linear | 10 | 6.3 pp | 12 |
| [FACTS Multimodal → NYT Connections](https://benchgap.net/calibration/facts-multimodal/current/to/nyt-connections/current.md) | Hill | 22 | 6.3 pp | 1 |
| [MaxxBench → DAYJOB](https://benchgap.net/calibration/maxxbench/current/to/dayjob/current.md) | linear | 8 | 6.3 pp | 1 |
| [AA-SciCode → SWE-bench Verified](https://benchgap.net/calibration/aa-sci-code/current/to/swe-bench-verified/current.md) | linear | 35 | 6.3 pp | 23 |
| [cursorBench41 → FrontierCode 2.0](https://benchgap.net/calibration/cursor-bench41/current/to/frontiercode-2/current.md) | inverse Michaelis–Menten | 10 | 6.3 pp | 0 |
| [PerceptionBench → ZeroBench](https://benchgap.net/calibration/perceptionbench/current/to/zerobench/current.md) | Hill | 5 | 6.3 pp | 0 |
| [MKA → AA-AnalystAgent](https://benchgap.net/calibration/mka/current/to/aa-analyst-agent/current.md) | linear | 12 | 6.3 pp | 1 |
| [Chess Puzzles → DAYJOB](https://benchgap.net/calibration/chesspuzzles/current/to/dayjob/current.md) | Hill | 19 | 6.3 pp | 17 |
| [Mystery Game Puzzles → SkillsBench official leaderboard (BenchFlow)](https://benchgap.net/calibration/mysterygamepuzzles/current/to/skillsbench/current.md) | linear | 14 | 6.3 pp | 6 |
| [AA-SciCode → SWE-bench Pro](https://benchgap.net/calibration/aa-sci-code/current/to/swe-bench-pro/current.md) | inverse Michaelis–Menten | 43 | 6.3 pp | 16 |
| [Decision Index Language → Decision Index Arts](https://benchgap.net/calibration/jevdecisionindex03language/current/to/jevdecisionindex03arts/current.md) | linear | 110 | 6.3 pp | 0 |
| [PostTrainBench v1.1 → Bug Hunt Bench](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1-1/current/to/bug-hunt-bench/current.md) | Hill | 7 | 6.3 pp | 8 |
| [ExploitBench Internal Port → StealthMark](https://benchgap.net/calibration/exploitbenchinternalport/current/to/stealthmark/current.md) | linear | 6 | 6.3 pp | 0 |
| [We-Math → CharXiv](https://benchgap.net/calibration/wemath/current/to/charxiv/current.md) | Michaelis–Menten | 5 | 6.3 pp | 0 |
| [APEX-Agents → HLE w/ tools](https://benchgap.net/calibration/apexagents/current/to/hlewithtools/current.md) | Michaelis–Menten | 6 | 6.3 pp | 1 |
| [OSWorld-Verified → ApprenticeBench](https://benchgap.net/calibration/osworld-verified/current/to/apprenticebench/current.md) | linear | 7 | 6.3 pp | 10 |
| [MKA → GDPval-AA](https://benchgap.net/calibration/mka/current/to/gdpvalaanormalized/current.md) | offset logistic | 15 | 6.4 pp | 2 |
| [MMMLU → AA-GPQA Diamond](https://benchgap.net/calibration/mmmlu/current/to/aa-gpqa-diamond/current.md) | linear | 9 | 6.4 pp | 0 |
| [Terminal-Bench 2.0 → Terminal-Bench 2.1 (Vals AI)](https://benchgap.net/calibration/terminal-bench%23terminal-bench-2-0/current/to/vals-terminal-bench21/current.md) | linear | 19 | 6.4 pp | 16 |
| [Artificial Analysis Decision Index v1.0 → Decision Index Full Score](https://benchgap.net/calibration/aa-decision-index/current/to/decisionindex/current.md) | linear | 11 | 6.4 pp | 0 |
| [ResearchClawBench → AA Agentic Index](https://benchgap.net/calibration/researchclawbench/current/to/aa-agentic-index/current.md) | inverse Michaelis–Menten | 11 | 6.4 pp | 3 |
| [BlenderWorld (text-only) → NYT Connections](https://benchgap.net/calibration/blenderworld-text-only/current/to/nyt-connections/current.md) | Hill | 5 | 6.4 pp | 2 |
| [Claw-Eval → MCP-Tasks](https://benchgap.net/calibration/claw-eval/current/to/mcptasks/current.md) | inverse Michaelis–Menten | 5 | 6.4 pp | 13 |
| [ApprenticeBench → GDPval-AA](https://benchgap.net/calibration/apprenticebench/current/to/gdpvalaanormalized/current.md) | Michaelis–Menten | 16 | 6.4 pp | 0 |
| [GDPval-AA → MKA](https://benchgap.net/calibration/gdpvalaanormalized/current/to/mka/current.md) | Michaelis–Menten | 15 | 6.4 pp | 31 |
| [BrowseComp → MazeBench](https://benchgap.net/calibration/browsecomp/current/to/mazebench/current.md) | inverse Michaelis–Menten | 14 | 6.4 pp | 21 |
| [Hermes Index → ExploitGym](https://benchgap.net/calibration/hermesindex/current/to/exploitgym/current.md) | Hill | 6 | 6.4 pp | 2 |
| [BrowseComp → Claw-Eval](https://benchgap.net/calibration/browsecomp/current/to/claw-eval/current.md) | offset logistic | 15 | 6.4 pp | 18 |
| [AA Coding Index → DeepSWE](https://benchgap.net/calibration/aa-coding-index/current/to/deepswe/current.md) | Hill | 20 | 6.4 pp | 16 |
| [BrowseComp → CWE-Bench-AA](https://benchgap.net/calibration/browsecomp/current/to/aa-cwe-bench/current.md) | Michaelis–Menten | 8 | 6.4 pp | 26 |
| [MMMU-Pro → ERQA](https://benchgap.net/calibration/mmmu-pro/current/to/erqa/current.md) | offset logistic | 15 | 6.4 pp | 7 |
| [ExtractBench Pro → ParseBench-2](https://benchgap.net/calibration/extractbench-pro/current/to/parsebench/current.md) | linear | 5 | 6.4 pp | 1 |
| [ApprenticeBench → AA Agentic Index](https://benchgap.net/calibration/apprenticebench/current/to/aa-agentic-index/current.md) | linear | 15 | 6.4 pp | 0 |
| [AA-AnalystAgent → MazeBench](https://benchgap.net/calibration/aa-analyst-agent/current/to/mazebench/current.md) | offset logistic | 22 | 6.4 pp | 1 |
| [GDPval-AA → CWE-bench v1](https://benchgap.net/calibration/gdpvalaanormalized/current/to/cwe-bench%23cwe-bench-v1/current.md) | Michaelis–Menten | 10 | 6.4 pp | 86 |
| [DRACO → Hermes Bench](https://benchgap.net/calibration/draco/current/to/hermesbench/current.md) | Michaelis–Menten | 5 | 6.4 pp | 1 |
| [DeepsecBench-AA → MazeBench](https://benchgap.net/calibration/aa-deepsec-bench/current/to/mazebench/current.md) | linear | 17 | 6.4 pp | 0 |
| [ExploitBench Internal Port → Mystery Game Puzzles](https://benchgap.net/calibration/exploitbenchinternalport/current/to/mysterygamepuzzles/current.md) | Michaelis–Menten | 7 | 6.4 pp | 2 |
| [DeepSWE → CursorBench 3.2](https://benchgap.net/calibration/deepswe/current/to/cursor-bench32/current.md) | Hill | 10 | 6.4 pp | 8 |
| [V* → MLVU (M-Avg)](https://benchgap.net/calibration/vstar/current/to/mlvuavg/current.md) | offset logistic | 6 | 6.4 pp | 0 |
| [CharXiv → MedXpertQA (MM)](https://benchgap.net/calibration/charxiv/current/to/medxpertqamm/current.md) | linear | 6 | 6.4 pp | 16 |
| [AA Coding Index → SWE-bench Verified](https://benchgap.net/calibration/aa-coding-index/current/to/swe-bench-verified/current.md) | Michaelis–Menten | 34 | 6.5 pp | 11 |
| [Vals MMLU-Pro → AA-Omniscience Accuracy](https://benchgap.net/calibration/vals-mmlu-pro/current/to/omniscienceaccuracy/current.md) | inverse Michaelis–Menten | 68 | 6.5 pp | 14 |
| [DAYJOB Healthcare → FACTS Multimodal](https://benchgap.net/calibration/dayjob-healthcare/current/to/facts-multimodal/current.md) | Hill | 15 | 6.5 pp | 0 |
| [HLE w/o tools → HLE](https://benchgap.net/calibration/hlenotools/current/to/hle/current.md) | Michaelis–Menten + offset | 36 | 6.5 pp | 16 |
| [ExploitBench → Furniture Assembly](https://benchgap.net/calibration/exploitbench/current/to/furnitureassembly/current.md) | linear | 9 | 6.5 pp | 7 |
| [ScreenSpot Pro → V*](https://benchgap.net/calibration/screenspot-pro/current/to/vstar/current.md) | Hill | 5 | 6.5 pp | 17 |
| [FrontierMath (legacy) → FrontierMath v2 (Tiers 1-3)](https://benchgap.net/calibration/frontiermath/current/to/frontiermathv2tiers13/current.md) | offset logistic | 7 | 6.5 pp | 0 |
| [Furniture Assembly → FACTS Multimodal](https://benchgap.net/calibration/furnitureassembly/current/to/facts-multimodal/current.md) | Hill | 18 | 6.5 pp | 0 |
| [WildGuardTest (Decision Index) → Decision Index Knowledge](https://benchgap.net/calibration/aa-decision-wild-guard-test/current/to/jevdecisionindex03knowledge/current.md) | Hill | 11 | 6.5 pp | 0 |
| [Vals GPQA Diamond → ARC-AGI-1](https://benchgap.net/calibration/vals-gpqa-diamond/current/to/arcagi1/current.md) | Hill | 28 | 6.5 pp | 30 |
| [React Native Evals → Vibe Code Bench](https://benchgap.net/calibration/reactnativeevals/current/to/vibecodebench/current.md) | offset logistic | 7 | 6.5 pp | 9 |
| [Bug Hunt Bench → NL2Repo](https://benchgap.net/calibration/bug-hunt-bench/current/to/nl2repo/current.md) | Michaelis–Menten | 5 | 6.5 pp | 6 |
| [DAYJOB → FACTS Multimodal](https://benchgap.net/calibration/dayjob/current/to/facts-multimodal/current.md) | linear | 15 | 6.5 pp | 0 |
| [Hermes Bench → GDPval-AA](https://benchgap.net/calibration/hermesbench/current/to/gdpvalaanormalized/current.md) | linear | 13 | 6.5 pp | 0 |
| [SuperGPQA → MMLU-Pro](https://benchgap.net/calibration/supergpqa/current/to/mmlu-pro/current.md) | Michaelis–Menten | 18 | 6.5 pp | 2 |
| [VulcanBench Frontier v4 → DeepSWE](https://benchgap.net/calibration/vulcanbench/current/to/deepswe/current.md) | Hill | 15 | 6.5 pp | 2 |
| [HLE → AA-HLE](https://benchgap.net/calibration/hle/current/to/aa-hle/current.md) | Michaelis–Menten + offset | 59 | 6.5 pp | 3 |
| [Video-MME (with subtitle) → VideoMMMU](https://benchgap.net/calibration/videommewithsub/current/to/videommmu/current.md) | inverse Michaelis–Menten | 11 | 6.5 pp | 0 |
| [CountBench → ERQA](https://benchgap.net/calibration/countbench/current/to/erqa/current.md) | Hill | 6 | 6.5 pp | 0 |
| [Furniture Assembly → ExploitBench](https://benchgap.net/calibration/furnitureassembly/current/to/exploitbench/current.md) | offset logistic | 9 | 6.5 pp | 20 |
| [Terminal-Bench 4.0 → Terminal-Bench 2.1](https://benchgap.net/calibration/terminal-bench/current/to/terminal-bench%23terminal-bench-2-1/current.md) | offset logistic | 10 | 6.5 pp | 6 |
| [AA-MMMU-Pro → MedXpertQA (MM)](https://benchgap.net/calibration/aa-mmmu-pro/current/to/medxpertqamm/current.md) | offset logistic | 6 | 6.5 pp | 45 |
| [HealthBench Professional (raw, Artificial Analysis) → HLE w/o tools](https://benchgap.net/calibration/aa-health-bench-professional-raw/current/to/hlenotools/current.md) | Hill | 5 | 6.5 pp | 0 |
| [MMLU-ProX → MAXIFE](https://benchgap.net/calibration/mmluprox/current/to/maxife/current.md) | inverse Michaelis–Menten | 5 | 6.5 pp | 5 |
| [Furniture Assembly → StealthMark](https://benchgap.net/calibration/furnitureassembly/current/to/stealthmark/current.md) | Hill | 21 | 6.5 pp | 11 |
| [ApprenticeBench → τ²-bench results](https://benchgap.net/calibration/apprenticebench/current/to/tau2-bench/current.md) | Michaelis–Menten | 8 | 6.5 pp | 6 |
| [Vals GPQA Diamond → HLE w/o tools](https://benchgap.net/calibration/vals-gpqa-diamond/current/to/hlenotools/current.md) | offset logistic | 21 | 6.5 pp | 0 |
| [Vibe Code Bench → SWE-Rebench](https://benchgap.net/calibration/vibecodebench/current/to/swe-rebench/current.md) | Hill | 5 | 6.5 pp | 27 |
| [Hermes Bench → MazeBench](https://benchgap.net/calibration/hermesbench/current/to/mazebench/current.md) | offset logistic | 13 | 6.5 pp | 0 |
| [GPQA → HLE](https://benchgap.net/calibration/gpqa/current/to/hle/current.md) | inverse Michaelis–Menten | 53 | 6.5 pp | 60 |
| [SEC-Bench Pro → ExploitBench Internal Port](https://benchgap.net/calibration/secbenchpro/current/to/exploitbenchinternalport/current.md) | inverse Michaelis–Menten | 9 | 6.5 pp | 2 |
| [Artificial Analysis Decision Index v1.0 → Decision Index Retrieval](https://benchgap.net/calibration/aa-decision-index/current/to/jevdecisionindex03retrieval/current.md) | Michaelis–Menten | 11 | 6.5 pp | 2 |
| [Gert Labs → BrowseComp](https://benchgap.net/calibration/gertlabs/current/to/browsecomp/current.md) | linear | 15 | 6.6 pp | 15 |
| [AA-IFBench → IFBench](https://benchgap.net/calibration/aa-if-bench/current/to/ifbench/current.md) | Hill | 13 | 6.6 pp | 127 |
| [SWE-bench Verified → Vals SWE-bench](https://benchgap.net/calibration/swe-bench-verified/current/to/valsswebench/current.md) | linear | 26 | 6.6 pp | 49 |
| [SEC-Bench Pro → StealthMark](https://benchgap.net/calibration/secbenchpro/current/to/stealthmark/current.md) | linear | 11 | 6.6 pp | 0 |
| [Terminal-Bench 3.0 → Terminal-Bench 2.1](https://benchgap.net/calibration/terminal-bench%23terminal-bench-3-0/current/to/terminal-bench%23terminal-bench-2-1/current.md) | linear | 11 | 6.6 pp | 0 |
| [AA LiveCodeBench → Vibe Code Bench](https://benchgap.net/calibration/aa-live-code-bench/current/to/vibecodebench/current.md) | offset logistic | 7 | 6.6 pp | 3 |
| [Vals MMLU-Pro → Vals GPQA Diamond](https://benchgap.net/calibration/vals-mmlu-pro/current/to/vals-gpqa-diamond/current.md) | Hill | 85 | 6.6 pp | 0 |
| [WildGuardTest (Decision Index) → Decision Index Retrieval](https://benchgap.net/calibration/aa-decision-wild-guard-test/current/to/jevdecisionindex03retrieval/current.md) | offset logistic | 11 | 6.6 pp | 0 |
| [MKA → MazeBench](https://benchgap.net/calibration/mka/current/to/mazebench/current.md) | linear | 18 | 6.6 pp | 0 |
| [Finance Agent v2 → MazeBench](https://benchgap.net/calibration/financeagentv2/current/to/mazebench/current.md) | linear | 6 | 6.6 pp | 1 |
| [WideResearch → OSWorld-Verified](https://benchgap.net/calibration/wideresearch/current/to/osworld-verified/current.md) | Hill | 5 | 6.6 pp | 0 |
| [Agents' Last Exam → MKA](https://benchgap.net/calibration/agentslastexam/current/to/mka/current.md) | Michaelis–Menten + offset | 8 | 6.6 pp | 1 |
| [DAYJOB → ExploitBench Internal Port](https://benchgap.net/calibration/dayjob/current/to/exploitbenchinternalport/current.md) | inverse Michaelis–Menten | 5 | 6.6 pp | 0 |
| [AA MMLU-Pro → AA-HLE](https://benchgap.net/calibration/aa-mmlu-pro/current/to/aa-hle/current.md) | linear | 13 | 6.6 pp | 0 |
| [APEX-Agents → MCP Atlas](https://benchgap.net/calibration/apexagents/current/to/mcpatlas/current.md) | offset logistic | 6 | 6.6 pp | 6 |
| [Vals GPQA Diamond → MMLU-Pro](https://benchgap.net/calibration/vals-gpqa-diamond/current/to/mmlu-pro/current.md) | Michaelis–Menten | 10 | 6.6 pp | 0 |
| [C-Eval → AA-HLE](https://benchgap.net/calibration/ceval/current/to/aa-hle/current.md) | linear | 7 | 6.6 pp | 0 |
| [Furniture Assembly → MaxxBench](https://benchgap.net/calibration/furnitureassembly/current/to/maxxbench/current.md) | linear | 10 | 6.6 pp | 5 |
| [CharXiv w/o tools → MMMU-Pro](https://benchgap.net/calibration/charxivnotools/current/to/mmmu-pro/current.md) | inverse Michaelis–Menten | 9 | 6.6 pp | 0 |
| [NYT Connections → OTelBench](https://benchgap.net/calibration/nyt-connections/current/to/otelbench/current.md) | offset logistic | 8 | 6.6 pp | 60 |
| [ARC-AGI-2 → HLE w/o tools](https://benchgap.net/calibration/arc-agi-2/current/to/hlenotools/current.md) | linear | 18 | 6.6 pp | 0 |
| [OSWorld-Verified → Gert Labs](https://benchgap.net/calibration/osworld-verified/current/to/gertlabs/current.md) | inverse Michaelis–Menten | 16 | 6.6 pp | 11 |
| [SWE-Atlas Codebase QnA → DeepSWE](https://benchgap.net/calibration/swe-atlas-codebase-qna/current/to/deepswe/current.md) | Michaelis–Menten + offset | 8 | 6.6 pp | 0 |
| [Video-MME (w/o subtitle) → We-Math](https://benchgap.net/calibration/videommenosub/current/to/wemath/current.md) | Michaelis–Menten | 6 | 6.6 pp | 0 |
| [AA-SciCode → FrontierCode 1.1 Main](https://benchgap.net/calibration/aa-sci-code/current/to/frontiercode/current.md) | offset logistic | 11 | 6.6 pp | 12 |
| [PostTrainBench → ScarfBench](https://benchgap.net/calibration/post-train-bench12/current/to/scarfbench/current.md) | Michaelis–Menten | 7 | 6.6 pp | 0 |
| [SkillsBench → MCP Atlas](https://benchgap.net/calibration/skills-bench/current/to/mcpatlas/current.md) | Michaelis–Menten | 5 | 6.6 pp | 2 |
| [When2Call (Decision Index) → Decision Index Retrieval](https://benchgap.net/calibration/aa-decision-when2-call/current/to/jevdecisionindex03retrieval/current.md) | Michaelis–Menten | 11 | 6.7 pp | 0 |
| [ARC-AGI-3 → CritPt](https://benchgap.net/calibration/arcagi3/current/to/critpt/current.md) | linear | 21 | 6.7 pp | 0 |
| [SpireClimb → FACTS Multimodal](https://benchgap.net/calibration/spireclimb/current/to/facts-multimodal/current.md) | Hill | 7 | 6.7 pp | 0 |
| [SWE-bench Pro → FrontierCode 2.0](https://benchgap.net/calibration/swe-bench-pro/current/to/frontiercode-2/current.md) | Hill | 8 | 6.7 pp | 62 |
| [BrowseComp → VITA-Bench](https://benchgap.net/calibration/browsecomp/current/to/vitabench/current.md) | Hill | 6 | 6.7 pp | 57 |
| [MCP Atlas → WideResearch](https://benchgap.net/calibration/mcpatlas/current/to/wideresearch/current.md) | offset logistic | 13 | 6.7 pp | 4 |
| [Bug Hunt Bench → FrontierCode 2.0](https://benchgap.net/calibration/bug-hunt-bench/current/to/frontiercode-2/current.md) | linear | 21 | 6.7 pp | 2 |
| [FACTS Multimodal → DAYJOB Healthcare](https://benchgap.net/calibration/facts-multimodal/current/to/dayjob-healthcare/current.md) | linear | 15 | 6.7 pp | 0 |
| [Decision Index Tools → Decision Index Arts](https://benchgap.net/calibration/jevdecisionindex03tools/current/to/jevdecisionindex03arts/current.md) | linear | 110 | 6.7 pp | 0 |
| [AA Agentic Index → HLE w/ tools](https://benchgap.net/calibration/aa-agentic-index/current/to/hlewithtools/current.md) | linear | 10 | 6.7 pp | 0 |
| [CWE-bench → MazeBench](https://benchgap.net/calibration/cwe-bench-v2/current/to/mazebench/current.md) | linear | 17 | 6.7 pp | 0 |
| [AA Tau3 Banking → GDPval-AA](https://benchgap.net/calibration/aa-tau3-banking/current/to/gdpvalaanormalized/current.md) | Michaelis–Menten | 17 | 6.7 pp | 1 |
| [SimpleVQA → We-Math](https://benchgap.net/calibration/simplevqa/current/to/wemath/current.md) | linear | 6 | 6.7 pp | 2 |
| [MMLU-Pro → AA-HLE](https://benchgap.net/calibration/mmlu-pro/current/to/aa-hle/current.md) | offset logistic | 50 | 6.7 pp | 1 |
| [APEX-Agents-AA → Toolathlon](https://benchgap.net/calibration/apexagentsaa/current/to/toolathlon/current.md) | linear | 15 | 6.7 pp | 3 |
| [SuperGPQA → AA-Omniscience Accuracy](https://benchgap.net/calibration/supergpqa/current/to/omniscienceaccuracy/current.md) | Michaelis–Menten | 12 | 6.7 pp | 7 |
| [AutomationBench → BrowseComp](https://benchgap.net/calibration/automationbench/current/to/browsecomp/current.md) | Michaelis–Menten | 10 | 6.7 pp | 3 |
| [HealthBench Professional → BioBasicsBench](https://benchgap.net/calibration/healthbenchprofessional/current/to/biobasicsbench/current.md) | offset logistic | 5 | 6.7 pp | 1 |
| [GPQA → ARC-AGI-1](https://benchgap.net/calibration/gpqa/current/to/arcagi1/current.md) | linear | 19 | 6.7 pp | 31 |
| [MMLU-Pro → AA-GPQA Diamond](https://benchgap.net/calibration/mmlu-pro/current/to/aa-gpqa-diamond/current.md) | Hill | 50 | 6.7 pp | 0 |
| [ScreenSpot Pro → ZeroBench](https://benchgap.net/calibration/screenspot-pro/current/to/zerobench/current.md) | Hill | 5 | 6.7 pp | 5 |
| [Terminal-Bench 2.1 (Vals AI) → Terminal-Bench 2.1](https://benchgap.net/calibration/vals-terminal-bench21/current/to/terminal-bench%23terminal-bench-2-1/current.md) | Hill | 31 | 6.7 pp | 25 |
| [BrowseComp → DeepSearchQA](https://benchgap.net/calibration/browsecomp/current/to/deepsearchqa/current.md) | Hill | 16 | 6.8 pp | 30 |
| [AA ITBench → MCP Atlas](https://benchgap.net/calibration/aa-itbench/current/to/mcpatlas/current.md) | Michaelis–Menten | 9 | 6.8 pp | 8 |
| [AA-SciCode → VulcanBench Frontier v4](https://benchgap.net/calibration/aa-sci-code/current/to/vulcanbench/current.md) | Michaelis–Menten | 18 | 6.8 pp | 50 |
| [Gert Labs → AA Agentic Index](https://benchgap.net/calibration/gertlabs/current/to/aa-agentic-index/current.md) | inverse Michaelis–Menten | 19 | 6.8 pp | 19 |
| [Gert Labs → Claw-Eval](https://benchgap.net/calibration/gertlabs/current/to/claw-eval/current.md) | Michaelis–Menten | 22 | 6.8 pp | 14 |
| [GDP.pdf → HLE w/ tools](https://benchgap.net/calibration/aa-gdp-pdf/current/to/hlewithtools/current.md) | linear | 6 | 6.8 pp | 6 |
| [StealthMark → ExploitBench Internal Port](https://benchgap.net/calibration/stealthmark/current/to/exploitbenchinternalport/current.md) | inverse Michaelis–Menten | 6 | 6.8 pp | 11 |
| [AA-HLE → HLE-Verified](https://benchgap.net/calibration/aa-hle/current/to/hleverified/current.md) | offset logistic | 6 | 6.8 pp | 24 |
| [MaxxBench → FrontierPhysics](https://benchgap.net/calibration/maxxbench/current/to/frontierphysics/current.md) | Hill | 9 | 6.8 pp | 0 |
| [AA-MMMU-Pro → OfficeQA Pro](https://benchgap.net/calibration/aa-mmmu-pro/current/to/officeqapro/current.md) | offset logistic | 9 | 6.8 pp | 67 |
| [ARC-AGI-3 Provider Adapter → ARC-AGI-3](https://benchgap.net/calibration/arc-agi-3-provider-adapter/current/to/arcagi3/current.md) | Hill | 13 | 6.8 pp | 0 |
| [FrontierCode 1.1 Main → VulcanBench Frontier v4](https://benchgap.net/calibration/frontiercode/current/to/vulcanbench/current.md) | offset logistic | 5 | 6.8 pp | 0 |
| [Bug Hunt Bench → cursorBench41](https://benchgap.net/calibration/bug-hunt-bench/current/to/cursor-bench41/current.md) | linear | 12 | 6.8 pp | 0 |
| [GPQA Diamond → HLE](https://benchgap.net/calibration/gpqa-diamond/current/to/hle/current.md) | offset logistic | 42 | 6.8 pp | 14 |
| [CharXiv → ScreenSpot Pro](https://benchgap.net/calibration/charxiv/current/to/screenspot-pro/current.md) | Hill | 14 | 6.8 pp | 25 |
| [CritPt → BioMysteryBench (human-difficult)](https://benchgap.net/calibration/critpt/current/to/biomysterybenchhumandifficult/current.md) | Michaelis–Menten | 6 | 6.8 pp | 201 |
| [When2Call (Decision Index) → Decision Index Knowledge](https://benchgap.net/calibration/aa-decision-when2-call/current/to/jevdecisionindex03knowledge/current.md) | Hill | 11 | 6.8 pp | 0 |
| [DeepSearchQA → BrowseComp](https://benchgap.net/calibration/deepsearchqa/current/to/browsecomp/current.md) | offset logistic | 16 | 6.8 pp | 1 |
| [MMLU-Redux → HLE](https://benchgap.net/calibration/mmluredux/current/to/hle/current.md) | linear | 9 | 6.8 pp | 0 |
| [CountBench → Video-MME (with subtitle)](https://benchgap.net/calibration/countbench/current/to/videommewithsub/current.md) | Hill | 7 | 6.8 pp | 0 |
| [FrontierCode 1.1 Main → SWE-bench Pro](https://benchgap.net/calibration/frontiercode/current/to/swe-bench-pro/current.md) | Hill | 9 | 6.8 pp | 3 |
| [CWE-Bench-AA → GDP.pdf](https://benchgap.net/calibration/aa-cwe-bench/current/to/aa-gdp-pdf/current.md) | offset logistic | 11 | 6.8 pp | 0 |
| [OSWorld 2.0 → AA Agentic Index](https://benchgap.net/calibration/osworld2/current/to/aa-agentic-index/current.md) | Hill | 16 | 6.8 pp | 0 |
| [ResearchClawBench → OSWorld 2.0](https://benchgap.net/calibration/researchclawbench/current/to/osworld2/current.md) | inverse Michaelis–Menten | 5 | 6.9 pp | 3 |
| [HLE-MultiChoice (Decision Index) → Decision Index Full Score](https://benchgap.net/calibration/aa-decision-hle-multi-choice/current/to/decisionindex/current.md) | offset logistic | 11 | 6.9 pp | 0 |
| [MMLU-Pro → Vals MMLU-Pro](https://benchgap.net/calibration/mmlu-pro/current/to/vals-mmlu-pro/current.md) | Michaelis–Menten | 9 | 6.9 pp | 1 |
| [Vibe Code Bench → AA Coding Index](https://benchgap.net/calibration/vibecodebench/current/to/aa-coding-index/current.md) | linear | 16 | 6.9 pp | 8 |
| [FrontierPhysics → SEC-Bench Pro](https://benchgap.net/calibration/frontierphysics/current/to/secbenchpro/current.md) | linear | 11 | 6.9 pp | 5 |
| [OmniDocBench 1.5 → We-Math](https://benchgap.net/calibration/omnidocbench15/current/to/wemath/current.md) | Hill | 5 | 6.9 pp | 0 |
| [SpireClimb → SEC-Bench Pro](https://benchgap.net/calibration/spireclimb/current/to/secbenchpro/current.md) | offset logistic | 6 | 6.9 pp | 1 |
| [Terminal-Bench 2.1 (Vals AI) → Terminal-Bench 3.0](https://benchgap.net/calibration/vals-terminal-bench21/current/to/terminal-bench%23terminal-bench-3-0/current.md) | Hill | 11 | 6.9 pp | 24 |
| [Bug Hunt Bench → DeepSWE](https://benchgap.net/calibration/bug-hunt-bench/current/to/deepswe/current.md) | Michaelis–Menten | 24 | 6.9 pp | 1 |
| [DynaMath → MMMU-Pro](https://benchgap.net/calibration/dynamath/current/to/mmmu-pro/current.md) | offset logistic | 7 | 6.9 pp | 0 |
| [DeepsecBench-AA → GDPval-AA](https://benchgap.net/calibration/aa-deepsec-bench/current/to/gdpvalaanormalized/current.md) | Michaelis–Menten + offset | 18 | 6.9 pp | 1 |
| [ResearchClawBench → Toolathlon](https://benchgap.net/calibration/researchclawbench/current/to/toolathlon/current.md) | offset logistic | 9 | 6.9 pp | 3 |
| [CursorBench 3.1 → SWE-bench Pro](https://benchgap.net/calibration/cursor-bench31/current/to/swe-bench-pro/current.md) | Michaelis–Menten + offset | 5 | 6.9 pp | 0 |
| [ParseBench-1 → MMMU-Pro](https://benchgap.net/calibration/parse-bench-llama-index/current/to/mmmu-pro/current.md) | linear | 15 | 6.9 pp | 5 |
| [MMLU → AA-GPQA Diamond](https://benchgap.net/calibration/mmlu/current/to/aa-gpqa-diamond/current.md) | offset logistic | 24 | 6.9 pp | 2 |
| [HLE-MultiChoice (Decision Index) → When2Call (Decision Index)](https://benchgap.net/calibration/aa-decision-hle-multi-choice/current/to/aa-decision-when2-call/current.md) | Hill | 19 | 6.9 pp | 0 |
| [OfficeQA → SWE-bench Multimodal](https://benchgap.net/calibration/officeqa/current/to/swe-bench-multimodal/current.md) | inverse Michaelis–Menten | 5 | 6.9 pp | 0 |
| [CWE-bench v1 → MazeBench](https://benchgap.net/calibration/cwe-bench%23cwe-bench-v1/current/to/mazebench/current.md) | Hill | 10 | 6.9 pp | 0 |
| [Decision Index Tools → When2Call (Decision Index)](https://benchgap.net/calibration/jevdecisionindex03tools/current/to/aa-decision-when2-call/current.md) | linear | 11 | 6.9 pp | 2 |
| [AA Terminal-Bench 4.0 → Terminal-Bench 2.1 (Vals AI)](https://benchgap.net/calibration/aa-terminal-bench4/current/to/vals-terminal-bench21/current.md) | linear | 7 | 6.9 pp | 15 |
| [Decision Index Retrieval → Decision Index Full Score](https://benchgap.net/calibration/jevdecisionindex03retrieval/current/to/decisionindex/current.md) | offset logistic | 110 | 6.9 pp | 0 |
| [Terminal-Bench-Science 0.1 → Terminal-Bench 4.0](https://benchgap.net/calibration/terminal-bench-science/current/to/terminal-bench/current.md) | Michaelis–Menten | 11 | 6.9 pp | 0 |
| [ARC-AGI-1 → HLE](https://benchgap.net/calibration/arcagi1/current/to/hle/current.md) | offset logistic | 16 | 7.0 pp | 1 |
| [SpireClimb → DAYJOB](https://benchgap.net/calibration/spireclimb/current/to/dayjob/current.md) | Michaelis–Menten | 6 | 7.0 pp | 0 |
| [CritPt → Vals GPQA Diamond](https://benchgap.net/calibration/critpt/current/to/vals-gpqa-diamond/current.md) | Hill | 70 | 7.0 pp | 1 |
| [AA-GPQA Diamond → MMLU-Pro](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/mmlu-pro/current.md) | Hill | 50 | 7.0 pp | 16 |
| [SimpleVQA → ERQA](https://benchgap.net/calibration/simplevqa/current/to/erqa/current.md) | linear | 12 | 7.0 pp | 0 |
| [Hermes Bench → CWE-bench v1](https://benchgap.net/calibration/hermesbench/current/to/cwe-bench%23cwe-bench-v1/current.md) | Hill | 5 | 7.0 pp | 0 |
| [Agents' Last Exam → GDP.pdf](https://benchgap.net/calibration/agentslastexam/current/to/aa-gdp-pdf/current.md) | Michaelis–Menten + offset | 5 | 7.0 pp | 0 |
| [When2Call (Decision Index) → Decision Index Full Score](https://benchgap.net/calibration/aa-decision-when2-call/current/to/decisionindex/current.md) | Michaelis–Menten | 11 | 7.0 pp | 0 |
| [FrontierPhysics → MaxxBench](https://benchgap.net/calibration/frontierphysics/current/to/maxxbench/current.md) | linear | 9 | 7.0 pp | 2 |
| [BabyVision w/ Python → ZeroBench w/ Python](https://benchgap.net/calibration/babyvisionpython/current/to/zerobenchpython/current.md) | Hill | 5 | 7.0 pp | 4 |
| [APEX-Agents-AA → τ²-bench results](https://benchgap.net/calibration/apexagentsaa/current/to/tau2-bench/current.md) | Michaelis–Menten | 21 | 7.0 pp | 5 |
| [ExploitBench Internal Port → ExploitBench](https://benchgap.net/calibration/exploitbenchinternalport/current/to/exploitbench/current.md) | Michaelis–Menten | 7 | 7.0 pp | 0 |
| [Artificial Analysis Decision Index v1.0 → Decision Index Language](https://benchgap.net/calibration/aa-decision-index/current/to/jevdecisionindex03language/current.md) | linear | 11 | 7.0 pp | 0 |
| [MLVU (M-Avg) → We-Math](https://benchgap.net/calibration/mlvuavg/current/to/wemath/current.md) | Michaelis–Menten | 6 | 7.0 pp | 0 |
| [MLVU (M-Avg) → VideoMMMU](https://benchgap.net/calibration/mlvuavg/current/to/videommmu/current.md) | inverse Michaelis–Menten | 10 | 7.0 pp | 0 |
| [React Native Evals → Vals SWE-bench](https://benchgap.net/calibration/reactnativeevals/current/to/valsswebench/current.md) | inverse Michaelis–Menten | 7 | 7.0 pp | 3 |
| [Agents' Last Exam → Hermes Index](https://benchgap.net/calibration/agentslastexam/current/to/hermesindex/current.md) | Hill | 7 | 7.0 pp | 3 |
| [WildGuardTest (Decision Index) → Decision Index Arts](https://benchgap.net/calibration/aa-decision-wild-guard-test/current/to/jevdecisionindex03arts/current.md) | offset logistic | 11 | 7.0 pp | 0 |
| [ParseBench-1 → OfficeQA Pro](https://benchgap.net/calibration/parse-bench-llama-index/current/to/officeqapro/current.md) | Michaelis–Menten | 9 | 7.0 pp | 6 |
| [Toolathlon → MCP-Tasks](https://benchgap.net/calibration/toolathlon/current/to/mcptasks/current.md) | Michaelis–Menten | 5 | 7.0 pp | 9 |
| [DAYJOB → Furniture Assembly](https://benchgap.net/calibration/dayjob/current/to/furnitureassembly/current.md) | Michaelis–Menten + offset | 16 | 7.1 pp | 3 |
| [VITA-Bench → Claw-Eval](https://benchgap.net/calibration/vitabench/current/to/claw-eval/current.md) | Hill | 7 | 7.1 pp | 0 |
| [HLE → AA-GPQA Diamond](https://benchgap.net/calibration/hle/current/to/aa-gpqa-diamond/current.md) | Hill | 51 | 7.1 pp | 3 |
| [HealthBench Hard → CritPt](https://benchgap.net/calibration/healthbench-hard/current/to/critpt/current.md) | Hill | 12 | 7.1 pp | 0 |
| [AA Coding Index → Vals LiveCodeBench](https://benchgap.net/calibration/aa-coding-index/current/to/valslivecodebench/current.md) | Michaelis–Menten | 49 | 7.1 pp | 9 |
| [ARC-AGI-1 → HLE w/o tools](https://benchgap.net/calibration/arcagi1/current/to/hlenotools/current.md) | Hill | 14 | 7.1 pp | 0 |
| [HLE w/ tools → BrowseComp](https://benchgap.net/calibration/hlewithtools/current/to/browsecomp/current.md) | Michaelis–Menten | 22 | 7.1 pp | 1 |
| [AA Tau3 Banking → AutomationBench](https://benchgap.net/calibration/aa-tau3-banking/current/to/automationbench/current.md) | inverse Michaelis–Menten | 7 | 7.1 pp | 10 |
| [MATH-500 → AIME 2024](https://benchgap.net/calibration/math-500/current/to/aime2024/current.md) | offset logistic | 15 | 7.1 pp | 4 |
| [Vibe Code Bench → Vals SWE-bench](https://benchgap.net/calibration/vibecodebench/current/to/valsswebench/current.md) | Michaelis–Menten + offset | 29 | 7.1 pp | 8 |
| [CharXiv w/o tools → ZeroBench](https://benchgap.net/calibration/charxivnotools/current/to/zerobench/current.md) | Hill | 5 | 7.1 pp | 7 |
| [Decision Index Retrieval → Decision Index Arts](https://benchgap.net/calibration/jevdecisionindex03retrieval/current/to/jevdecisionindex03arts/current.md) | linear | 110 | 7.1 pp | 0 |
| [AA-Omniscience Accuracy → HLE-Verified](https://benchgap.net/calibration/omniscienceaccuracy/current/to/hleverified/current.md) | inverse Michaelis–Menten | 6 | 7.1 pp | 48 |
| [FrontierPhysics → FACTS Multimodal](https://benchgap.net/calibration/frontierphysics/current/to/facts-multimodal/current.md) | linear | 15 | 7.1 pp | 1 |
| [SimpleVQA → MedXpertQA (MM)](https://benchgap.net/calibration/simplevqa/current/to/medxpertqamm/current.md) | offset logistic | 6 | 7.1 pp | 1 |
| [DeepSearchQA → OSWorld-Verified](https://benchgap.net/calibration/deepsearchqa/current/to/osworld-verified/current.md) | Michaelis–Menten | 7 | 7.1 pp | 2 |
| [OmniDocBench 1.5 → SimpleVQA](https://benchgap.net/calibration/omnidocbench15/current/to/simplevqa/current.md) | Hill | 8 | 7.1 pp | 2 |
| [MMMU → SimpleVQA](https://benchgap.net/calibration/mmmu/current/to/simplevqa/current.md) | inverse Michaelis–Menten | 9 | 7.1 pp | 0 |
| [HealthBench Professional (raw, Artificial Analysis) → PhysicianBench](https://benchgap.net/calibration/aa-health-bench-professional-raw/current/to/physicianbench/current.md) | Hill | 5 | 7.1 pp | 0 |
| [BrowseComp → Gert Labs](https://benchgap.net/calibration/browsecomp/current/to/gertlabs/current.md) | inverse Michaelis–Menten | 15 | 7.1 pp | 29 |
| [ProgramBench → PostTrainBench v1.1](https://benchgap.net/calibration/programbench/current/to/posttrainbench%23posttrainbench-v1-1/current.md) | Michaelis–Menten + offset | 5 | 7.2 pp | 0 |
| [Gert Labs → Toolathlon](https://benchgap.net/calibration/gertlabs/current/to/toolathlon/current.md) | Michaelis–Menten + offset | 13 | 7.2 pp | 16 |
| [MMMU → We-Math](https://benchgap.net/calibration/mmmu/current/to/wemath/current.md) | Michaelis–Menten + offset | 6 | 7.2 pp | 16 |
| [Mystery Game Puzzles → ExploitBench](https://benchgap.net/calibration/mysterygamepuzzles/current/to/exploitbench/current.md) | Michaelis–Menten + offset | 10 | 7.2 pp | 33 |
| [BioBasicsBench → AA-Omniscience Accuracy](https://benchgap.net/calibration/biobasicsbench/current/to/omniscienceaccuracy/current.md) | Michaelis–Menten | 7 | 7.2 pp | 0 |
| [SimpleVQA → MathVision](https://benchgap.net/calibration/simplevqa/current/to/mathvision/current.md) | Michaelis–Menten | 9 | 7.2 pp | 0 |
| [MaxxBench → DAYJOB Healthcare](https://benchgap.net/calibration/maxxbench/current/to/dayjob-healthcare/current.md) | linear | 8 | 7.2 pp | 0 |
| [Vals LiveCodeBench → CursorBench 3.2](https://benchgap.net/calibration/valslivecodebench/current/to/cursor-bench32/current.md) | offset logistic | 16 | 7.2 pp | 18 |
| [HLE w/ tools → OSWorld 2.1](https://benchgap.net/calibration/hlewithtools/current/to/osworld-2-1/current.md) | Michaelis–Menten + offset | 8 | 7.2 pp | 11 |
| [Toolathlon → Gert Labs](https://benchgap.net/calibration/toolathlon/current/to/gertlabs/current.md) | Hill | 13 | 7.2 pp | 1 |
| [Furniture Assembly → ExploitBench Internal Port](https://benchgap.net/calibration/furnitureassembly/current/to/exploitbenchinternalport/current.md) | Hill | 7 | 7.2 pp | 0 |
| [ScreenSpot Pro → Video-MME (with subtitle)](https://benchgap.net/calibration/screenspot-pro/current/to/videommewithsub/current.md) | linear | 6 | 7.2 pp | 4 |
| [NL2Repo → Bug Hunt Bench](https://benchgap.net/calibration/nl2repo/current/to/bug-hunt-bench/current.md) | inverse Michaelis–Menten | 5 | 7.2 pp | 28 |
| [Decision Index Tools → Decision Index Language](https://benchgap.net/calibration/jevdecisionindex03tools/current/to/jevdecisionindex03language/current.md) | linear | 110 | 7.2 pp | 0 |
| [AA-GPQA Diamond → BioBasicsBench](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/biobasicsbench/current.md) | inverse Michaelis–Menten | 5 | 7.2 pp | 0 |
| [Vals SWE-bench → DeepSWE](https://benchgap.net/calibration/valsswebench/current/to/deepswe/current.md) | offset logistic | 18 | 7.2 pp | 27 |
| [Hermes Bench → Hermes Index](https://benchgap.net/calibration/hermesbench/current/to/hermesindex/current.md) | inverse Michaelis–Menten | 16 | 7.2 pp | 0 |
| [CritPt → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/critpt/current/to/artificialanalysis/current.md) | Michaelis–Menten + offset | 154 | 7.2 pp | 31 |
| [HLE → SuperGPQA](https://benchgap.net/calibration/hle/current/to/supergpqa/current.md) | linear | 13 | 7.2 pp | 63 |
| [SWE-bench Pro → Vals SWE-bench](https://benchgap.net/calibration/swe-bench-pro/current/to/valsswebench/current.md) | Hill | 38 | 7.2 pp | 3 |
| [HLE w/o tools → ARC-AGI-3](https://benchgap.net/calibration/hlenotools/current/to/arcagi3/current.md) | Hill | 11 | 7.2 pp | 45 |
| [CyberGym-E2E-AA → AA EnterpriseOps-Gym](https://benchgap.net/calibration/aa-cyber-gym-e2e/current/to/aa-enterprise-ops-gym/current.md) | linear | 6 | 7.2 pp | 13 |
| [MKA → ExploitGym](https://benchgap.net/calibration/mka/current/to/exploitgym/current.md) | linear | 8 | 7.2 pp | 2 |
| [Decision Index Arts → Decision Index Language](https://benchgap.net/calibration/jevdecisionindex03arts/current/to/jevdecisionindex03language/current.md) | Michaelis–Menten | 110 | 7.2 pp | 0 |
| [HLE-Verified → Vals GPQA Diamond](https://benchgap.net/calibration/hleverified/current/to/vals-gpqa-diamond/current.md) | linear | 6 | 7.2 pp | 0 |
| [HLE-MultiChoice (Decision Index) → Decision Index Retrieval](https://benchgap.net/calibration/aa-decision-hle-multi-choice/current/to/jevdecisionindex03retrieval/current.md) | Michaelis–Menten | 11 | 7.2 pp | 0 |
| [NYT Connections → SpireClimb](https://benchgap.net/calibration/nyt-connections/current/to/spireclimb/current.md) | offset logistic | 10 | 7.2 pp | 57 |
| [OpenHarmony Bench v1.0 → Vals SWE-bench](https://benchgap.net/calibration/openharmonybench%23openharmonybench-v1-0/current/to/valsswebench/current.md) | linear | 11 | 7.3 pp | 0 |
| [VulcanBench Frontier v4 → cursorBench41](https://benchgap.net/calibration/vulcanbench/current/to/cursor-bench41/current.md) | inverse Michaelis–Menten | 6 | 7.3 pp | 0 |
| [AA-HLE → ARC-AGI-3 Provider Adapter](https://benchgap.net/calibration/aa-hle/current/to/arc-agi-3-provider-adapter/current.md) | offset logistic | 10 | 7.3 pp | 193 |
| [BrowseComp → CyberGym](https://benchgap.net/calibration/browsecomp/current/to/cybergym/current.md) | inverse Michaelis–Menten | 15 | 7.3 pp | 20 |
| [OSWorld 2.0 → HLE w/ tools](https://benchgap.net/calibration/osworld2/current/to/hlewithtools/current.md) | Michaelis–Menten | 5 | 7.3 pp | 0 |
| [AA-Omniscience Accuracy → CritPt](https://benchgap.net/calibration/omniscienceaccuracy/current/to/critpt/current.md) | inverse Michaelis–Menten | 159 | 7.3 pp | 0 |
| [MCP Atlas → DeepSearchQA](https://benchgap.net/calibration/mcpatlas/current/to/deepsearchqa/current.md) | offset logistic | 14 | 7.3 pp | 8 |
| [Terminal-Bench 2.1 → Terminal-Bench 2.1 (Vals AI)](https://benchgap.net/calibration/terminal-bench%23terminal-bench-2-1/current/to/vals-terminal-bench21/current.md) | linear | 31 | 7.3 pp | 52 |
| [FACTS Multimodal → ExploitBench](https://benchgap.net/calibration/facts-multimodal/current/to/exploitbench/current.md) | Hill | 6 | 7.3 pp | 2 |
| [Toolathlon → JobBench](https://benchgap.net/calibration/toolathlon/current/to/jobbench/current.md) | Michaelis–Menten | 5 | 7.3 pp | 9 |
| [MathVision → SimpleVQA](https://benchgap.net/calibration/mathvision/current/to/simplevqa/current.md) | linear | 9 | 7.3 pp | 11 |
| [Decision Index Full Score → Decision Index Retrieval](https://benchgap.net/calibration/decisionindex/current/to/jevdecisionindex03retrieval/current.md) | Michaelis–Menten | 110 | 7.3 pp | 0 |
| [Artificial Analysis Intelligence Index → AA-Omniscience Accuracy](https://benchgap.net/calibration/artificialanalysis/current/to/omniscienceaccuracy/current.md) | Michaelis–Menten + offset | 163 | 7.3 pp | 12 |
| [ERQA → MathVision](https://benchgap.net/calibration/erqa/current/to/mathvision/current.md) | Michaelis–Menten | 12 | 7.3 pp | 0 |
| [CritPt → HealthBench Professional (length-adjusted, Artificial Analysis)](https://benchgap.net/calibration/critpt/current/to/aa-health-bench-professional/current.md) | Michaelis–Menten + offset | 6 | 7.3 pp | 8 |
| [BrowseComp → Toolathlon-Verified](https://benchgap.net/calibration/browsecomp/current/to/toolathlonverified/current.md) | linear | 11 | 7.4 pp | 10 |
| [HLE-MultiChoice (Decision Index) → Decision Index Language](https://benchgap.net/calibration/aa-decision-hle-multi-choice/current/to/jevdecisionindex03language/current.md) | Michaelis–Menten | 11 | 7.4 pp | 0 |
| [MKA → OSWorld 2.1 Offline](https://benchgap.net/calibration/mka/current/to/osworld-2-1-offline/current.md) | Michaelis–Menten | 9 | 7.4 pp | 1 |
| [HealthBench (raw, Artificial Analysis) → PhysicianBench](https://benchgap.net/calibration/aa-health-bench/current/to/physicianbench/current.md) | inverse Michaelis–Menten | 5 | 7.4 pp | 1 |
| [MMLU-Redux → GPQA Diamond](https://benchgap.net/calibration/mmluredux/current/to/gpqa-diamond/current.md) | linear | 9 | 7.4 pp | 0 |
| [FACTS Multimodal → DAYJOB](https://benchgap.net/calibration/facts-multimodal/current/to/dayjob/current.md) | linear | 15 | 7.4 pp | 0 |
| [AA-SciCode → NL2Repo](https://benchgap.net/calibration/aa-sci-code/current/to/nl2repo/current.md) | linear | 16 | 7.4 pp | 11 |
| [APEX-Agents → AutomationBench](https://benchgap.net/calibration/apexagents/current/to/automationbench/current.md) | Michaelis–Menten | 6 | 7.4 pp | 5 |
| [AutomationBench → APEX-Agents](https://benchgap.net/calibration/automationbench/current/to/apexagents/current.md) | offset logistic | 6 | 7.4 pp | 26 |
| [IFBench → AA-IFBench](https://benchgap.net/calibration/ifbench/current/to/aa-if-bench/current.md) | offset logistic | 13 | 7.4 pp | 44 |
| [Vals GPQA Diamond → AA-Omniscience Accuracy](https://benchgap.net/calibration/vals-gpqa-diamond/current/to/omniscienceaccuracy/current.md) | Hill | 68 | 7.4 pp | 0 |
| [MMMU-Pro → We-Math](https://benchgap.net/calibration/mmmu-pro/current/to/wemath/current.md) | Michaelis–Menten + offset | 6 | 7.4 pp | 23 |
| [ExploitGym → GDP.pdf](https://benchgap.net/calibration/exploitgym/current/to/aa-gdp-pdf/current.md) | Michaelis–Menten | 6 | 7.4 pp | 5 |
| [HealthBench Professional (raw, Artificial Analysis) → AA-HLE](https://benchgap.net/calibration/aa-health-bench-professional-raw/current/to/aa-hle/current.md) | Hill | 6 | 7.4 pp | 0 |
| [ARC-AGI-2 → AA-Omniscience Accuracy](https://benchgap.net/calibration/arc-agi-2/current/to/omniscienceaccuracy/current.md) | Hill | 40 | 7.4 pp | 4 |
| [EBR-bench → Furniture Assembly](https://benchgap.net/calibration/ebrbench/current/to/furnitureassembly/current.md) | Michaelis–Menten + offset | 27 | 7.4 pp | 5 |
| [MMLU-Redux → GPQA](https://benchgap.net/calibration/mmluredux/current/to/gpqa/current.md) | linear | 16 | 7.4 pp | 0 |
| [JobBench → Toolathlon](https://benchgap.net/calibration/jobbench/current/to/toolathlon/current.md) | Hill | 5 | 7.4 pp | 4 |
| [ARC-AGI-3 → AA-HLE](https://benchgap.net/calibration/arcagi3/current/to/aa-hle/current.md) | linear | 22 | 7.4 pp | 0 |
| [GPQA → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/gpqa/current/to/artificialanalysis/current.md) | offset logistic | 77 | 7.4 pp | 25 |
| [AA-HLE → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/aa-hle/current/to/artificialanalysis/current.md) | offset logistic | 167 | 7.4 pp | 1 |
| [CharXiv w/o tools → BabyVision](https://benchgap.net/calibration/charxivnotools/current/to/babyvision/current.md) | inverse Michaelis–Menten | 5 | 7.4 pp | 9 |
| [AA-SciCode → PostTrainBench v1.1](https://benchgap.net/calibration/aa-sci-code/current/to/posttrainbench%23posttrainbench-v1-1/current.md) | Hill | 13 | 7.4 pp | 18 |
| [ARC-AGI-3 Provider Adapter → AA-HLE](https://benchgap.net/calibration/arc-agi-3-provider-adapter/current/to/aa-hle/current.md) | linear | 10 | 7.4 pp | 0 |
| [Toolathlon-Verified → MCP Atlas](https://benchgap.net/calibration/toolathlonverified/current/to/mcpatlas/current.md) | Michaelis–Menten | 12 | 7.5 pp | 4 |
| [V* → ScreenSpot Pro](https://benchgap.net/calibration/vstar/current/to/screenspot-pro/current.md) | Michaelis–Menten + offset | 5 | 7.5 pp | 1 |
| [ExploitBench → Mystery Game Puzzles](https://benchgap.net/calibration/exploitbench/current/to/mysterygamepuzzles/current.md) | inverse Michaelis–Menten | 10 | 7.5 pp | 6 |
| [APEX-Agents → GDPval-AA](https://benchgap.net/calibration/apexagents/current/to/gdpvalaanormalized/current.md) | linear | 7 | 7.5 pp | 1 |
| [Decision Index Knowledge → Decision Index Tools](https://benchgap.net/calibration/jevdecisionindex03knowledge/current/to/jevdecisionindex03tools/current.md) | Hill | 110 | 7.5 pp | 0 |
| [Video-MME (w/o subtitle) → SimpleVQA](https://benchgap.net/calibration/videommenosub/current/to/simplevqa/current.md) | inverse Michaelis–Menten | 7 | 7.5 pp | 0 |
| [EBR-bench → SEC-Bench Pro](https://benchgap.net/calibration/ebrbench/current/to/secbenchpro/current.md) | Michaelis–Menten + offset | 9 | 7.5 pp | 13 |
| [AA-SciCode → DeepSWE](https://benchgap.net/calibration/aa-sci-code/current/to/deepswe/current.md) | Hill | 40 | 7.5 pp | 33 |
| [AA AutomationBench → DeepsecBench-AA](https://benchgap.net/calibration/aa-automation-bench/current/to/aa-deepsec-bench/current.md) | linear | 11 | 7.5 pp | 12 |
| [SWE-bench Pro → AA Coding Index](https://benchgap.net/calibration/swe-bench-pro/current/to/aa-coding-index/current.md) | Hill | 38 | 7.5 pp | 2 |
| [ApprenticeBench → DeepSearchQA](https://benchgap.net/calibration/apprenticebench/current/to/deepsearchqa/current.md) | linear | 5 | 7.5 pp | 1 |
| [MMMU-Pro → SimpleVQA](https://benchgap.net/calibration/mmmu-pro/current/to/simplevqa/current.md) | Hill | 16 | 7.5 pp | 25 |
| [AA AutomationBench → HLE w/ tools](https://benchgap.net/calibration/aa-automation-bench/current/to/hlewithtools/current.md) | linear | 6 | 7.5 pp | 0 |
| [AA-Omniscience Accuracy → HLE w/o tools](https://benchgap.net/calibration/omniscienceaccuracy/current/to/hlenotools/current.md) | Michaelis–Menten | 37 | 7.5 pp | 0 |
| [AA-MMMU-Pro → V*](https://benchgap.net/calibration/aa-mmmu-pro/current/to/vstar/current.md) | inverse Michaelis–Menten | 10 | 7.5 pp | 63 |
| [MazeBench → HLE w/ tools](https://benchgap.net/calibration/mazebench/current/to/hlewithtools/current.md) | offset logistic | 15 | 7.5 pp | 3 |
| [MCP Atlas → Toolathlon](https://benchgap.net/calibration/mcpatlas/current/to/toolathlon/current.md) | Hill | 17 | 7.5 pp | 4 |
| [MKA → DeepsecBench-AA](https://benchgap.net/calibration/mka/current/to/aa-deepsec-bench/current.md) | Michaelis–Menten | 11 | 7.5 pp | 2 |
| [FACTS Multimodal → ExploitBench Internal Port](https://benchgap.net/calibration/facts-multimodal/current/to/exploitbenchinternalport/current.md) | linear | 6 | 7.6 pp | 0 |
| [Decision Index Tools → Decision Index Retrieval](https://benchgap.net/calibration/jevdecisionindex03tools/current/to/jevdecisionindex03retrieval/current.md) | linear | 110 | 7.6 pp | 0 |
| [ARC-AGI-3 → AA-Omniscience Accuracy](https://benchgap.net/calibration/arcagi3/current/to/omniscienceaccuracy/current.md) | linear | 24 | 7.6 pp | 0 |
| [MMLU → GPQA](https://benchgap.net/calibration/mmlu/current/to/gpqa/current.md) | offset logistic | 26 | 7.6 pp | 2 |
| [SuperGPQA → AA-HLE](https://benchgap.net/calibration/supergpqa/current/to/aa-hle/current.md) | offset logistic | 17 | 7.6 pp | 2 |
| [ScreenSpot Pro → MLVU (M-Avg)](https://benchgap.net/calibration/screenspot-pro/current/to/mlvuavg/current.md) | linear | 6 | 7.6 pp | 5 |
| [ExploitGym → GDPval-AA](https://benchgap.net/calibration/exploitgym/current/to/gdpvalaanormalized/current.md) | Michaelis–Menten + offset | 17 | 7.6 pp | 4 |
| [DeepSWE → SciCode](https://benchgap.net/calibration/deepswe/current/to/scicode/current.md) | Michaelis–Menten + offset | 5 | 7.6 pp | 14 |
| [Toolathlon → APEX-Agents-AA](https://benchgap.net/calibration/toolathlon/current/to/apexagentsaa/current.md) | Hill | 15 | 7.6 pp | 0 |
| [ARC-AGI-1 → HealthBench (raw)](https://benchgap.net/calibration/arcagi1/current/to/healthbench/current.md) | offset logistic | 6 | 7.6 pp | 4 |
| [DRACO → MCP Atlas](https://benchgap.net/calibration/draco/current/to/mcpatlas/current.md) | Michaelis–Menten | 6 | 7.6 pp | 1 |
| [OCRBench V2 → MMMU-Pro](https://benchgap.net/calibration/ocrbenchv2/current/to/mmmu-pro/current.md) | inverse Michaelis–Menten | 8 | 7.6 pp | 1 |
| [HLE w/ tools → GDPval-AA](https://benchgap.net/calibration/hlewithtools/current/to/gdpvalaanormalized/current.md) | Hill | 19 | 7.7 pp | 5 |
| [CritPt → HLE w/o tools](https://benchgap.net/calibration/critpt/current/to/hlenotools/current.md) | Hill | 40 | 7.7 pp | 1 |
| [MaxxBench → StealthMark](https://benchgap.net/calibration/maxxbench/current/to/stealthmark/current.md) | inverse Michaelis–Menten | 9 | 7.7 pp | 3 |
| [AA Harvey LAB v1.0 → BrowseComp](https://benchgap.net/calibration/aa-harvey-lab/current/to/browsecomp/current.md) | inverse Michaelis–Menten | 5 | 7.7 pp | 0 |
| [FailureSensorIQ → OTelBench](https://benchgap.net/calibration/failuresensoriq/current/to/otelbench/current.md) | offset logistic | 5 | 7.7 pp | 7 |
| [FrontierSWE v2 → Bug Hunt Bench](https://benchgap.net/calibration/frontierswev2/current/to/bug-hunt-bench/current.md) | offset logistic | 20 | 7.7 pp | 4 |
| [DeepSWE → SWE-bench Pro](https://benchgap.net/calibration/deepswe/current/to/swe-bench-pro/current.md) | offset logistic | 25 | 7.7 pp | 8 |
| [OSWorld 2.0 → GDPval-AA](https://benchgap.net/calibration/osworld2/current/to/gdpvalaanormalized/current.md) | Michaelis–Menten | 22 | 7.7 pp | 0 |
| [HLE w/o tools → AA-Omniscience Accuracy](https://benchgap.net/calibration/hlenotools/current/to/omniscienceaccuracy/current.md) | Hill | 37 | 7.7 pp | 14 |
| [PostTrainBench v1.1 → FrontierCode 2.0](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1-1/current/to/frontiercode-2/current.md) | inverse Michaelis–Menten | 9 | 7.7 pp | 1 |
| [When2Call (Decision Index) → Decision Index Language](https://benchgap.net/calibration/aa-decision-when2-call/current/to/jevdecisionindex03language/current.md) | Michaelis–Menten | 11 | 7.7 pp | 0 |
| [AA-GPQA Diamond → AA-Omniscience Accuracy](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/omniscienceaccuracy/current.md) | offset logistic | 136 | 7.7 pp | 27 |
| [MMAnswerBench → FrontierMath v2 (Tiers 1-3)](https://benchgap.net/calibration/mmanswerbench/current/to/frontiermathv2tiers13/current.md) | linear | 7 | 7.8 pp | 3 |
| [CritPt → ARC-AGI-1](https://benchgap.net/calibration/critpt/current/to/arcagi1/current.md) | linear | 37 | 7.8 pp | 88 |
| [PostTrainBench → CursorBench 4.0](https://benchgap.net/calibration/post-train-bench12/current/to/cursor-bench40/current.md) | Hill | 8 | 7.8 pp | 0 |
| [MCP Atlas → BrowseComp](https://benchgap.net/calibration/mcpatlas/current/to/browsecomp/current.md) | Hill | 35 | 7.8 pp | 2 |
| [Artificial Analysis Intelligence Index → GPQA](https://benchgap.net/calibration/artificialanalysis/current/to/gpqa/current.md) | Hill | 77 | 7.8 pp | 8 |
| [AA-Omniscience Accuracy → BioMysteryBench (human-difficult)](https://benchgap.net/calibration/omniscienceaccuracy/current/to/biomysterybenchhumandifficult/current.md) | Michaelis–Menten | 6 | 7.8 pp | 11 |
| [AA Agentic Index → MCP Atlas](https://benchgap.net/calibration/aa-agentic-index/current/to/mcpatlas/current.md) | offset logistic | 23 | 7.8 pp | 30 |
| [MCP Atlas → Gert Labs](https://benchgap.net/calibration/mcpatlas/current/to/gertlabs/current.md) | linear | 14 | 7.8 pp | 6 |
| [MaxxBench → FACTS Multimodal](https://benchgap.net/calibration/maxxbench/current/to/facts-multimodal/current.md) | Michaelis–Menten | 7 | 7.8 pp | 0 |
| [GDPval-AA → WideResearch](https://benchgap.net/calibration/gdpvalaanormalized/current/to/wideresearch/current.md) | Michaelis–Menten | 6 | 7.8 pp | 54 |
| [EBR-bench → SkillsBench official leaderboard (BenchFlow)](https://benchgap.net/calibration/ebrbench/current/to/skillsbench/current.md) | linear | 10 | 7.8 pp | 0 |
| [BrowseComp → GDPval-AA](https://benchgap.net/calibration/browsecomp/current/to/gdpvalaanormalized/current.md) | Hill | 34 | 7.8 pp | 10 |
| [AA-GPQA Diamond → HLE w/o tools](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/hlenotools/current.md) | linear | 25 | 7.8 pp | 0 |
| [AA-GPQA Diamond → SimpleQA](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/simpleqa/current.md) | linear | 9 | 7.8 pp | 78 |
| [OSWorld-Verified → APEX-Agents-AA](https://benchgap.net/calibration/osworld-verified/current/to/apexagentsaa/current.md) | offset logistic | 9 | 7.8 pp | 7 |
| [OSWorld 2.0 → CyberGym](https://benchgap.net/calibration/osworld2/current/to/cybergym/current.md) | Michaelis–Menten | 8 | 7.9 pp | 0 |
| [DeepsecBench-AA → CWE-Bench-AA](https://benchgap.net/calibration/aa-deepsec-bench/current/to/aa-cwe-bench/current.md) | Michaelis–Menten | 19 | 7.9 pp | 0 |
| [HLE w/ tools → AA AutomationBench](https://benchgap.net/calibration/hlewithtools/current/to/aa-automation-bench/current.md) | offset logistic | 6 | 7.9 pp | 25 |
| [AA-Omniscience Accuracy → ARC-AGI-1](https://benchgap.net/calibration/omniscienceaccuracy/current/to/arcagi1/current.md) | Michaelis–Menten | 39 | 7.9 pp | 4 |
| [AA EnterpriseOps-Gym → DeepSearchQA](https://benchgap.net/calibration/aa-enterprise-ops-gym/current/to/deepsearchqa/current.md) | inverse Michaelis–Menten | 5 | 7.9 pp | 1 |
| [CyberGym-E2E-AA → AA Cyber Index](https://benchgap.net/calibration/aa-cyber-gym-e2e/current/to/aa-cyber-index/current.md) | Hill | 21 | 7.9 pp | 0 |
| [MazeBench → CWE-bench](https://benchgap.net/calibration/mazebench/current/to/cwe-bench-v2/current.md) | linear | 17 | 7.9 pp | 18 |
| [MMMU → ScreenSpot Pro](https://benchgap.net/calibration/mmmu/current/to/screenspot-pro/current.md) | offset logistic | 7 | 7.9 pp | 16 |
| [EBR-bench → ExploitBench](https://benchgap.net/calibration/ebrbench/current/to/exploitbench/current.md) | Michaelis–Menten | 6 | 7.9 pp | 0 |
| [ScarfBench → AA-SciCode](https://benchgap.net/calibration/scarfbench/current/to/aa-sci-code/current.md) | Hill | 10 | 7.9 pp | 0 |
| [AA-AnalystAgent → OSWorld 2.1 Offline](https://benchgap.net/calibration/aa-analyst-agent/current/to/osworld-2-1-offline/current.md) | inverse Michaelis–Menten | 6 | 7.9 pp | 3 |
| [MCP Atlas → JobBench](https://benchgap.net/calibration/mcpatlas/current/to/jobbench/current.md) | inverse Michaelis–Menten | 13 | 7.9 pp | 21 |
| [CharXiv → We-Math](https://benchgap.net/calibration/charxiv/current/to/wemath/current.md) | offset logistic | 5 | 7.9 pp | 13 |
| [MirrorCode → FrontierPhysics](https://benchgap.net/calibration/mirrorcode/current/to/frontierphysics/current.md) | Michaelis–Menten | 21 | 7.9 pp | 0 |
| [Artificial Analysis Intelligence Index → ARC-AGI-1](https://benchgap.net/calibration/artificialanalysis/current/to/arcagi1/current.md) | linear | 39 | 7.9 pp | 14 |
| [ScreenSpot Pro → AA-MMMU-Pro](https://benchgap.net/calibration/screenspot-pro/current/to/aa-mmmu-pro/current.md) | linear | 13 | 7.9 pp | 4 |
| [CursorBench 4.0 → Bug Hunt Bench](https://benchgap.net/calibration/cursor-bench40/current/to/bug-hunt-bench/current.md) | Hill | 17 | 7.9 pp | 4 |
| [OSWorld-Verified → DeepSearchQA](https://benchgap.net/calibration/osworld-verified/current/to/deepsearchqa/current.md) | Hill | 7 | 7.9 pp | 9 |
| [Vals LiveCodeBench → NL2Repo](https://benchgap.net/calibration/valslivecodebench/current/to/nl2repo/current.md) | offset logistic | 13 | 7.9 pp | 9 |
| [ARC-AGI-3 → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/arcagi3/current/to/artificialanalysis/current.md) | linear | 24 | 8.0 pp | 0 |
| [OSWorld-Verified → AA Agentic Index](https://benchgap.net/calibration/osworld-verified/current/to/aa-agentic-index/current.md) | inverse Michaelis–Menten | 17 | 8.0 pp | 4 |
| [FrontierCode 2.0 → FrontierSWE v2](https://benchgap.net/calibration/frontiercode-2/current/to/frontierswev2/current.md) | Michaelis–Menten | 15 | 8.0 pp | 15 |
| [AA Harvey LAB v1.0 → AA Agentic Index](https://benchgap.net/calibration/aa-harvey-lab/current/to/aa-agentic-index/current.md) | Hill | 7 | 8.0 pp | 0 |
| [Toolathlon-Verified → AA-AnalystAgent](https://benchgap.net/calibration/toolathlonverified/current/to/aa-analyst-agent/current.md) | Hill | 6 | 8.0 pp | 1 |
| [MMMU → OCRBench V2](https://benchgap.net/calibration/mmmu/current/to/ocrbenchv2/current.md) | Michaelis–Menten | 5 | 8.0 pp | 11 |
| [VITA-Bench → BrowseComp](https://benchgap.net/calibration/vitabench/current/to/browsecomp/current.md) | linear | 6 | 8.0 pp | 0 |
| [ExploitGym → AA Agentic Index](https://benchgap.net/calibration/exploitgym/current/to/aa-agentic-index/current.md) | linear | 7 | 8.0 pp | 5 |
| [ExploitBench → Chess Puzzles](https://benchgap.net/calibration/exploitbench/current/to/chesspuzzles/current.md) | Hill | 9 | 8.0 pp | 6 |
| [Cookie-Bench → Chess Puzzles](https://benchgap.net/calibration/cookie-bench/current/to/chesspuzzles/current.md) | linear | 5 | 8.0 pp | 2 |
| [Video-MME (with subtitle) → MMMU](https://benchgap.net/calibration/videommewithsub/current/to/mmmu/current.md) | Michaelis–Menten | 8 | 8.0 pp | 0 |
| [ResearchClawBench → GDPval-AA](https://benchgap.net/calibration/researchclawbench/current/to/gdpvalaanormalized/current.md) | Michaelis–Menten | 14 | 8.0 pp | 3 |
| [AA-GPQA Diamond → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/artificialanalysis/current.md) | offset logistic | 147 | 8.0 pp | 0 |
| [ARC-AGI-1 → AA-Omniscience Accuracy](https://benchgap.net/calibration/arcagi1/current/to/omniscienceaccuracy/current.md) | Hill | 39 | 8.0 pp | 0 |
| [SEC-Bench Pro → ExploitBench](https://benchgap.net/calibration/secbenchpro/current/to/exploitbench/current.md) | linear | 9 | 8.0 pp | 0 |
| [AA-HLE → ARC-AGI-1](https://benchgap.net/calibration/aa-hle/current/to/arcagi1/current.md) | linear | 36 | 8.1 pp | 2 |
| [Vals MMLU-Pro → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/vals-mmlu-pro/current/to/artificialanalysis/current.md) | offset logistic | 72 | 8.1 pp | 0 |
| [ScarfBench → Vals SWE-bench](https://benchgap.net/calibration/scarfbench/current/to/valsswebench/current.md) | linear | 5 | 8.1 pp | 0 |
| [SimpleVQA → Video-MME (w/o subtitle)](https://benchgap.net/calibration/simplevqa/current/to/videommenosub/current.md) | Michaelis–Menten | 7 | 8.1 pp | 0 |
| [VulcanBench Frontier v4 → SWE-bench Pro](https://benchgap.net/calibration/vulcanbench/current/to/swe-bench-pro/current.md) | offset logistic | 9 | 8.1 pp | 1 |
| [AA-HLE → BioMysteryBench (human-difficult)](https://benchgap.net/calibration/aa-hle/current/to/biomysterybenchhumandifficult/current.md) | Michaelis–Menten | 6 | 8.1 pp | 3 |
| [RoboDojo Sim → AutomationBench](https://benchgap.net/calibration/robodojosim/current/to/automationbench/current.md) | Michaelis–Menten | 5 | 8.1 pp | 0 |
| [BrowseComp → AA Agentic Index](https://benchgap.net/calibration/browsecomp/current/to/aa-agentic-index/current.md) | inverse Michaelis–Menten | 21 | 8.2 pp | 11 |
| [QwenClawBench → VITA-Bench](https://benchgap.net/calibration/qwenclawbench/current/to/vitabench/current.md) | Michaelis–Menten | 6 | 8.2 pp | 2 |
| [APEX-Agents-AA → JobBench](https://benchgap.net/calibration/apexagentsaa/current/to/jobbench/current.md) | linear | 5 | 8.2 pp | 11 |
| [JobBench → DeepSearchQA](https://benchgap.net/calibration/jobbench/current/to/deepsearchqa/current.md) | linear | 9 | 8.2 pp | 13 |
| [ExploitGym → CWE-bench](https://benchgap.net/calibration/exploitgym/current/to/cwe-bench-v2/current.md) | Michaelis–Menten | 7 | 8.2 pp | 6 |
| [AA-Omniscience Accuracy → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/omniscienceaccuracy/current/to/artificialanalysis/current.md) | linear | 163 | 8.2 pp | 0 |
| [CharXiv w/o tools → LVBench](https://benchgap.net/calibration/charxivnotools/current/to/lvbench/current.md) | Hill | 6 | 8.2 pp | 12 |
| [AA-SciCode → Vals SWE-bench](https://benchgap.net/calibration/aa-sci-code/current/to/valsswebench/current.md) | linear | 42 | 8.2 pp | 15 |
| [MMLU → Vals GPQA Diamond](https://benchgap.net/calibration/mmlu/current/to/vals-gpqa-diamond/current.md) | Hill | 8 | 8.2 pp | 3 |
| [ExploitBench → SpireClimb](https://benchgap.net/calibration/exploitbench/current/to/spireclimb/current.md) | linear | 5 | 8.2 pp | 3 |
| [Decision Index Full Score → When2Call (Decision Index)](https://benchgap.net/calibration/decisionindex/current/to/aa-decision-when2-call/current.md) | Michaelis–Menten | 11 | 8.2 pp | 4 |
| [CWE-Bench-AA → AA AutomationBench](https://benchgap.net/calibration/aa-cwe-bench/current/to/aa-automation-bench/current.md) | offset logistic | 11 | 8.2 pp | 1 |
| [Terminal-Bench 4.1 → Terminal-Bench 2.1 (Vals AI)](https://benchgap.net/calibration/terminal-bench41/current/to/vals-terminal-bench21/current.md) | Michaelis–Menten + offset | 32 | 8.3 pp | 3 |
| [cursorBench41 → FrontierSWE v2](https://benchgap.net/calibration/cursor-bench41/current/to/frontierswev2/current.md) | linear | 12 | 8.3 pp | 0 |
| [HLE → ARC-AGI-3](https://benchgap.net/calibration/hle/current/to/arcagi3/current.md) | linear | 8 | 8.3 pp | 21 |
| [SWE-bench Pro → VulcanBench Frontier v4](https://benchgap.net/calibration/swe-bench-pro/current/to/vulcanbench/current.md) | Michaelis–Menten | 9 | 8.3 pp | 29 |
| [ARC-AGI-2 → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/arc-agi-2/current/to/artificialanalysis/current.md) | offset logistic | 40 | 8.3 pp | 1 |
| [ERQA → MMMU-Pro](https://benchgap.net/calibration/erqa/current/to/mmmu-pro/current.md) | Michaelis–Menten + offset | 15 | 8.3 pp | 0 |
| [APEX-Agents → Toolathlon-Verified](https://benchgap.net/calibration/apexagents/current/to/toolathlonverified/current.md) | offset logistic | 5 | 8.3 pp | 2 |
| [Artificial Analysis Intelligence Index → BioMysteryBench (human-difficult)](https://benchgap.net/calibration/artificialanalysis/current/to/biomysterybenchhumandifficult/current.md) | Michaelis–Menten | 6 | 8.3 pp | 16 |
| [SpireClimb → Chess Puzzles](https://benchgap.net/calibration/spireclimb/current/to/chesspuzzles/current.md) | Michaelis–Menten + offset | 11 | 8.3 pp | 0 |
| [AA Agentic Index → Toolathlon](https://benchgap.net/calibration/aa-agentic-index/current/to/toolathlon/current.md) | Hill | 15 | 8.3 pp | 0 |
| [AA Agentic Index → Gert Labs](https://benchgap.net/calibration/aa-agentic-index/current/to/gertlabs/current.md) | linear | 19 | 8.3 pp | 18 |
| [SWE-bench Pro → FrontierSWE v2](https://benchgap.net/calibration/swe-bench-pro/current/to/frontierswev2/current.md) | linear | 9 | 8.3 pp | 68 |
| [HealthBench Professional → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/healthbenchprofessional/current/to/artificialanalysis/current.md) | offset logistic | 19 | 8.3 pp | 1 |
| [Chartography (no tools) → SWE-bench Multimodal](https://benchgap.net/calibration/chartography/current/to/swe-bench-multimodal/current.md) | offset logistic | 5 | 8.3 pp | 6 |
| [AA-GPQA Diamond → MMLU](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/mmlu/current.md) | Hill | 24 | 8.3 pp | 17 |
| [OmniDocBench 1.5 → ERQA](https://benchgap.net/calibration/omnidocbench15/current/to/erqa/current.md) | Hill | 7 | 8.3 pp | 0 |
| [Decision Index Knowledge → Decision Index Retrieval](https://benchgap.net/calibration/jevdecisionindex03knowledge/current/to/jevdecisionindex03retrieval/current.md) | Michaelis–Menten + offset | 110 | 8.3 pp | 0 |
| [Vibe Code Bench → PostTrainBench](https://benchgap.net/calibration/vibecodebench/current/to/post-train-bench12/current.md) | Hill | 5 | 8.3 pp | 10 |
| [τ³-bench results → DeepPlanning](https://benchgap.net/calibration/tau3-bench/current/to/deepplanning/current.md) | offset logistic | 6 | 8.3 pp | 7 |
| [Decision Index Arts → Decision Index Retrieval](https://benchgap.net/calibration/jevdecisionindex03arts/current/to/jevdecisionindex03retrieval/current.md) | Michaelis–Menten + offset | 110 | 8.4 pp | 0 |
| [JobBench → Gert Labs](https://benchgap.net/calibration/jobbench/current/to/gertlabs/current.md) | Hill | 15 | 8.4 pp | 7 |
| [MMMU-Pro → ParseBench-1](https://benchgap.net/calibration/mmmu-pro/current/to/parse-bench-llama-index/current.md) | Hill | 15 | 8.4 pp | 14 |
| [GDPval-AA → OSWorld-Verified](https://benchgap.net/calibration/gdpvalaanormalized/current/to/osworld-verified/current.md) | linear | 21 | 8.4 pp | 58 |
| [SpireClimb → DAYJOB Healthcare](https://benchgap.net/calibration/spireclimb/current/to/dayjob-healthcare/current.md) | Michaelis–Menten + offset | 6 | 8.4 pp | 0 |
| [GPQA Diamond → AA-Omniscience Accuracy](https://benchgap.net/calibration/gpqa-diamond/current/to/omniscienceaccuracy/current.md) | inverse Michaelis–Menten | 46 | 8.4 pp | 16 |
| [Decision Index Language → Decision Index Tools](https://benchgap.net/calibration/jevdecisionindex03language/current/to/jevdecisionindex03tools/current.md) | Michaelis–Menten + offset | 110 | 8.4 pp | 0 |
| [CursorBench 3.2 → DeepSWE](https://benchgap.net/calibration/cursor-bench32/current/to/deepswe/current.md) | Michaelis–Menten | 10 | 8.4 pp | 0 |
| [Vals MMLU-Pro → ARC-AGI-1](https://benchgap.net/calibration/vals-mmlu-pro/current/to/arcagi1/current.md) | linear | 28 | 8.4 pp | 3 |
| [SciCode → DeepSWE](https://benchgap.net/calibration/scicode/current/to/deepswe/current.md) | inverse Michaelis–Menten | 5 | 8.4 pp | 12 |
| [Video-MME (with subtitle) → ScreenSpot Pro](https://benchgap.net/calibration/videommewithsub/current/to/screenspot-pro/current.md) | offset logistic | 6 | 8.4 pp | 1 |
| [MMLU-Pro (Arcee) → GPQA Diamond](https://benchgap.net/calibration/mmlu-pro-arcee/current/to/gpqa-diamond/current.md) | Michaelis–Menten | 6 | 8.4 pp | 0 |
| [FrontierCode 2.0 → SWE-bench Pro](https://benchgap.net/calibration/frontiercode-2/current/to/swe-bench-pro/current.md) | linear | 8 | 8.4 pp | 0 |
| [JobBench → OSWorld-Verified](https://benchgap.net/calibration/jobbench/current/to/osworld-verified/current.md) | Hill | 15 | 8.4 pp | 4 |
| [SWE-bench Verified → Vibe Code Bench](https://benchgap.net/calibration/swe-bench-verified/current/to/vibecodebench/current.md) | offset logistic | 11 | 8.4 pp | 88 |
| [DAYJOB Healthcare → MaxxBench](https://benchgap.net/calibration/dayjob-healthcare/current/to/maxxbench/current.md) | linear | 8 | 8.4 pp | 0 |
| [ARC-AGI-3 Provider Adapter → AA-Omniscience Accuracy](https://benchgap.net/calibration/arc-agi-3-provider-adapter/current/to/omniscienceaccuracy/current.md) | Michaelis–Menten + offset | 13 | 8.4 pp | 0 |
| [WildGuardTest (Decision Index) → When2Call (Decision Index)](https://benchgap.net/calibration/aa-decision-wild-guard-test/current/to/aa-decision-when2-call/current.md) | offset logistic | 19 | 8.4 pp | 0 |
| [τ²-bench results → BFCL v4](https://benchgap.net/calibration/tau2-bench/current/to/bfcl-v4/current.md) | Michaelis–Menten | 7 | 8.5 pp | 58 |
| [V* → ERQA](https://benchgap.net/calibration/vstar/current/to/erqa/current.md) | Michaelis–Menten | 6 | 8.5 pp | 0 |
| [AA-HLE → HLE](https://benchgap.net/calibration/aa-hle/current/to/hle/current.md) | offset logistic | 59 | 8.5 pp | 83 |
| [GDPval-AA → Gert Labs](https://benchgap.net/calibration/gdpvalaanormalized/current/to/gertlabs/current.md) | linear | 25 | 8.5 pp | 38 |
| [V* → We-Math](https://benchgap.net/calibration/vstar/current/to/wemath/current.md) | linear | 5 | 8.5 pp | 0 |
| [HLE-Verified → Vals MMLU-Pro](https://benchgap.net/calibration/hleverified/current/to/vals-mmlu-pro/current.md) | linear | 6 | 8.5 pp | 0 |
| [SkillsBench official leaderboard (BenchFlow) → OTelBench](https://benchgap.net/calibration/skillsbench/current/to/otelbench/current.md) | Hill | 5 | 8.5 pp | 0 |
| [FACTS Multimodal → FrontierPhysics](https://benchgap.net/calibration/facts-multimodal/current/to/frontierphysics/current.md) | Hill | 15 | 8.5 pp | 0 |
| [OSWorld-Verified → GDPval-AA](https://benchgap.net/calibration/osworld-verified/current/to/gdpvalaanormalized/current.md) | offset logistic | 21 | 8.5 pp | 4 |
| [SEC-Bench Pro → Furniture Assembly](https://benchgap.net/calibration/secbenchpro/current/to/furnitureassembly/current.md) | Hill | 11 | 8.5 pp | 0 |
| [GPQA → AA-Omniscience Accuracy](https://benchgap.net/calibration/gpqa/current/to/omniscienceaccuracy/current.md) | offset logistic | 72 | 8.5 pp | 15 |
| [GPQA → SuperGPQA](https://benchgap.net/calibration/gpqa/current/to/supergpqa/current.md) | Hill | 16 | 8.5 pp | 54 |
| [Decision Index Retrieval → When2Call (Decision Index)](https://benchgap.net/calibration/jevdecisionindex03retrieval/current/to/aa-decision-when2-call/current.md) | Michaelis–Menten + offset | 11 | 8.5 pp | 12 |
| [SpireClimb → NYT Connections](https://benchgap.net/calibration/spireclimb/current/to/nyt-connections/current.md) | linear | 10 | 8.5 pp | 2 |
| [DeepPlanning → VITA-Bench](https://benchgap.net/calibration/deepplanning/current/to/vitabench/current.md) | Hill | 5 | 8.5 pp | 0 |
| [ResearchClawBench → Gert Labs](https://benchgap.net/calibration/researchclawbench/current/to/gertlabs/current.md) | Michaelis–Menten + offset | 16 | 8.5 pp | 1 |
| [Hermes Index → DeepsecBench-AA](https://benchgap.net/calibration/hermesindex/current/to/aa-deepsec-bench/current.md) | offset logistic | 10 | 8.6 pp | 1 |
| [CritPt → HealthBench Hard](https://benchgap.net/calibration/critpt/current/to/healthbench-hard/current.md) | Michaelis–Menten | 12 | 8.6 pp | 164 |
| [Hermes Index → GDPval-AA](https://benchgap.net/calibration/hermesindex/current/to/gdpvalaanormalized/current.md) | Michaelis–Menten | 13 | 8.6 pp | 0 |
| [DAYJOB Healthcare → Furniture Assembly](https://benchgap.net/calibration/dayjob-healthcare/current/to/furnitureassembly/current.md) | Michaelis–Menten + offset | 16 | 8.6 pp | 0 |
| [ExploitBench Internal Port → EBR-bench](https://benchgap.net/calibration/exploitbenchinternalport/current/to/ebrbench/current.md) | linear | 6 | 8.6 pp | 3 |
| [GDP.pdf → GDPval-AA](https://benchgap.net/calibration/aa-gdp-pdf/current/to/gdpvalaanormalized/current.md) | linear | 13 | 8.6 pp | 1 |
| [MMLU-Redux → AA-GPQA Diamond](https://benchgap.net/calibration/mmluredux/current/to/aa-gpqa-diamond/current.md) | Hill | 9 | 8.6 pp | 0 |
| [React Native Evals → AA-SciCode](https://benchgap.net/calibration/reactnativeevals/current/to/aa-sci-code/current.md) | Michaelis–Menten + offset | 6 | 8.6 pp | 3 |
| [CharXiv → ParseBench-1](https://benchgap.net/calibration/charxiv/current/to/parse-bench-llama-index/current.md) | inverse Michaelis–Menten | 12 | 8.6 pp | 7 |
| [AA Terminal-Bench 2.1 → Terminal-Bench 3.0](https://benchgap.net/calibration/aa-terminal-bench21/current/to/terminal-bench%23terminal-bench-3-0/current.md) | Hill | 9 | 8.6 pp | 2 |
| [Toolathlon → CyberGym](https://benchgap.net/calibration/toolathlon/current/to/cybergym/current.md) | offset logistic | 10 | 8.6 pp | 0 |
| [VulcanBench v3 → Bug Hunt Bench](https://benchgap.net/calibration/vulcanbench%23vulcanbench-v3/current/to/bug-hunt-bench/current.md) | Hill | 7 | 8.6 pp | 1 |
| [Hermes Bench → DeepsecBench-AA](https://benchgap.net/calibration/hermesbench/current/to/aa-deepsec-bench/current.md) | offset logistic | 10 | 8.6 pp | 0 |
| [CyberGym → JobBench](https://benchgap.net/calibration/cybergym/current/to/jobbench/current.md) | Michaelis–Menten | 12 | 8.6 pp | 8 |
| [Vals MMLU-Pro → ARC-AGI-3](https://benchgap.net/calibration/vals-mmlu-pro/current/to/arcagi3/current.md) | linear | 13 | 8.6 pp | 33 |
| [BioBasicsBench → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/biobasicsbench/current/to/artificialanalysis/current.md) | Michaelis–Menten | 7 | 8.6 pp | 0 |
| [AA-GPQA Diamond → ARC-AGI-1](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/arcagi1/current.md) | Hill | 28 | 8.6 pp | 0 |
| [JobBench → MCP Atlas](https://benchgap.net/calibration/jobbench/current/to/mcpatlas/current.md) | Michaelis–Menten | 13 | 8.6 pp | 15 |
| [GPQA Diamond → ARC-AGI-3](https://benchgap.net/calibration/gpqa-diamond/current/to/arcagi3/current.md) | inverse Michaelis–Menten | 14 | 8.6 pp | 13 |
| [GDPval-AA → JobBench](https://benchgap.net/calibration/gdpvalaanormalized/current/to/jobbench/current.md) | linear | 16 | 8.7 pp | 67 |
| [MirrorCode → StealthMark](https://benchgap.net/calibration/mirrorcode/current/to/stealthmark/current.md) | linear | 24 | 8.7 pp | 2 |
| [AI2D_TEST → AA-MMMU-Pro](https://benchgap.net/calibration/ai2dtest/current/to/aa-mmmu-pro/current.md) | Hill | 6 | 8.7 pp | 0 |
| [GDPval-AA → Hermes Index](https://benchgap.net/calibration/gdpvalaanormalized/current/to/hermesindex/current.md) | Michaelis–Menten + offset | 13 | 8.7 pp | 81 |
| [EBR-bench → Mystery Game Puzzles](https://benchgap.net/calibration/ebrbench/current/to/mysterygamepuzzles/current.md) | linear | 32 | 8.7 pp | 0 |
| [VideoMMMU → We-Math](https://benchgap.net/calibration/videommmu/current/to/wemath/current.md) | Michaelis–Menten + offset | 6 | 8.7 pp | 0 |
| [NL2Repo → DeepSWE](https://benchgap.net/calibration/nl2repo/current/to/deepswe/current.md) | inverse Michaelis–Menten | 15 | 8.7 pp | 9 |
| [VideoMMMU → SimpleVQA](https://benchgap.net/calibration/videommmu/current/to/simplevqa/current.md) | Michaelis–Menten | 11 | 8.7 pp | 0 |
| [ScreenSpot Pro → MMMU-Pro](https://benchgap.net/calibration/screenspot-pro/current/to/mmmu-pro/current.md) | Michaelis–Menten + offset | 15 | 8.7 pp | 4 |
| [ScreenSpot Pro → ERQA](https://benchgap.net/calibration/screenspot-pro/current/to/erqa/current.md) | Michaelis–Menten | 10 | 8.7 pp | 4 |
| [IFEval → AA-IFBench](https://benchgap.net/calibration/ifeval/current/to/aa-if-bench/current.md) | inverse Michaelis–Menten | 31 | 8.8 pp | 22 |
| [VulcanBench Frontier v4 → FrontierCode 2.0](https://benchgap.net/calibration/vulcanbench/current/to/frontiercode-2/current.md) | linear | 14 | 8.8 pp | 1 |
| [GDP.pdf → CWE-bench](https://benchgap.net/calibration/aa-gdp-pdf/current/to/cwe-bench-v2/current.md) | Michaelis–Menten + offset | 14 | 8.8 pp | 1 |
| [HealthBench Professional (length-adjusted, Artificial Analysis) → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/aa-health-bench-professional/current/to/artificialanalysis/current.md) | linear | 6 | 8.8 pp | 0 |
| [Mystery Game Puzzles → MaxxBench](https://benchgap.net/calibration/mysterygamepuzzles/current/to/maxxbench/current.md) | Hill | 13 | 8.8 pp | 20 |
| [AIME26 → FrontierMath v2 (Tiers 1-3)](https://benchgap.net/calibration/aime2026/current/to/frontiermathv2tiers13/current.md) | linear | 9 | 8.8 pp | 16 |
| [MCP Atlas → Toolathlon-Verified](https://benchgap.net/calibration/mcpatlas/current/to/toolathlonverified/current.md) | linear | 12 | 8.8 pp | 2 |
| [OfficeQA Pro → SWE-bench Multimodal](https://benchgap.net/calibration/officeqapro/current/to/swe-bench-multimodal/current.md) | Hill | 6 | 8.8 pp | 4 |
| [Gert Labs → GDPval-AA](https://benchgap.net/calibration/gertlabs/current/to/gdpvalaanormalized/current.md) | linear | 25 | 8.8 pp | 9 |
| [OTelBench → BinaryAudit](https://benchgap.net/calibration/otelbench/current/to/binaryaudit/current.md) | offset logistic | 11 | 8.8 pp | 2 |
| [DeepsecBench-AA → CWE-bench](https://benchgap.net/calibration/aa-deepsec-bench/current/to/cwe-bench-v2/current.md) | linear | 9 | 8.9 pp | 0 |
| [ExploitBench Internal Port → SEC-Bench Pro](https://benchgap.net/calibration/exploitbenchinternalport/current/to/secbenchpro/current.md) | Michaelis–Menten + offset | 9 | 8.9 pp | 0 |
| [SEC-Bench Pro → NYT Connections](https://benchgap.net/calibration/secbenchpro/current/to/nyt-connections/current.md) | inverse Michaelis–Menten | 10 | 8.9 pp | 4 |
| [FrontierCode 2.0 → Bug Hunt Bench](https://benchgap.net/calibration/frontiercode-2/current/to/bug-hunt-bench/current.md) | Michaelis–Menten | 21 | 8.9 pp | 7 |
| [We-Math → Video-MME (w/o subtitle)](https://benchgap.net/calibration/wemath/current/to/videommenosub/current.md) | Michaelis–Menten + offset | 6 | 8.9 pp | 0 |
| [GDPval-AA → DeepsecBench-AA](https://benchgap.net/calibration/gdpvalaanormalized/current/to/aa-deepsec-bench/current.md) | Hill | 18 | 8.9 pp | 101 |
| [MMMLU → MMLU-Pro](https://benchgap.net/calibration/mmmlu/current/to/mmlu-pro/current.md) | Michaelis–Menten | 5 | 8.9 pp | 0 |
| [IMOAnswerBench → HMMT Feb 2026](https://benchgap.net/calibration/imoanswerbench/current/to/hmmtfeb2026/current.md) | linear | 9 | 8.9 pp | 1 |
| [AA Coding Index → SWE-Rebench](https://benchgap.net/calibration/aa-coding-index/current/to/swe-rebench/current.md) | offset logistic | 5 | 8.9 pp | 14 |
| [AA Cyber Index → HLE w/ tools](https://benchgap.net/calibration/aa-cyber-index/current/to/hlewithtools/current.md) | Michaelis–Menten | 8 | 8.9 pp | 0 |
| [EBR-bench → StealthMark](https://benchgap.net/calibration/ebrbench/current/to/stealthmark/current.md) | linear | 12 | 8.9 pp | 3 |
| [AA ITBench → AA Agentic Index](https://benchgap.net/calibration/aa-itbench/current/to/aa-agentic-index/current.md) | Hill | 17 | 8.9 pp | 0 |
| [FailureSensorIQ → BinaryAudit](https://benchgap.net/calibration/failuresensoriq/current/to/binaryaudit/current.md) | linear | 5 | 8.9 pp | 12 |
| [HLE w/o tools → MedXpertQA (Text)](https://benchgap.net/calibration/hlenotools/current/to/medxpertqatext/current.md) | Michaelis–Menten | 5 | 8.9 pp | 17 |
| [FrontierCode 1.1 Extended → SWE-bench Pro](https://benchgap.net/calibration/frontiercode11extended/current/to/swe-bench-pro/current.md) | linear | 6 | 9.0 pp | 0 |
| [ARC-AGI-1 → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/arcagi1/current/to/artificialanalysis/current.md) | Hill | 39 | 9.0 pp | 0 |
| [CWE-Bench-AA → GDPval-AA](https://benchgap.net/calibration/aa-cwe-bench/current/to/gdpvalaanormalized/current.md) | linear | 18 | 9.0 pp | 0 |
| [AA-Omniscience Accuracy → Vals GPQA Diamond](https://benchgap.net/calibration/omniscienceaccuracy/current/to/vals-gpqa-diamond/current.md) | Hill | 68 | 9.0 pp | 0 |
| [RealWorldQA → MMMU](https://benchgap.net/calibration/realworldqa/current/to/mmmu/current.md) | linear | 12 | 9.0 pp | 2 |
| [AA LiveCodeBench → Vals SWE-bench](https://benchgap.net/calibration/aa-live-code-bench/current/to/valsswebench/current.md) | Hill | 9 | 9.0 pp | 2 |
| [C-Eval → AA-GPQA Diamond](https://benchgap.net/calibration/ceval/current/to/aa-gpqa-diamond/current.md) | Hill | 7 | 9.0 pp | 0 |
| [BrowseComp → DeepsecBench-AA](https://benchgap.net/calibration/browsecomp/current/to/aa-deepsec-bench/current.md) | Hill | 8 | 9.0 pp | 19 |
| [FrontierCode 1.1 Main → CursorBench 4.0](https://benchgap.net/calibration/frontiercode/current/to/cursor-bench40/current.md) | inverse Michaelis–Menten | 5 | 9.0 pp | 0 |
| [AA Agentic Index → BrowseComp](https://benchgap.net/calibration/aa-agentic-index/current/to/browsecomp/current.md) | Michaelis–Menten | 21 | 9.0 pp | 14 |
| [AA Coding Index → Bug Hunt Bench](https://benchgap.net/calibration/aa-coding-index/current/to/bug-hunt-bench/current.md) | Hill | 10 | 9.0 pp | 15 |
| [AA ITBench → OSWorld-Verified](https://benchgap.net/calibration/aa-itbench/current/to/osworld-verified/current.md) | Michaelis–Menten + offset | 5 | 9.0 pp | 1 |
| [HMMT Feb 2026 → FrontierMath v2 (Tiers 1-3)](https://benchgap.net/calibration/hmmtfeb2026/current/to/frontiermathv2tiers13/current.md) | Hill | 11 | 9.0 pp | 2 |
| [FrontierPhysics → Furniture Assembly](https://benchgap.net/calibration/frontierphysics/current/to/furnitureassembly/current.md) | Michaelis–Menten + offset | 24 | 9.0 pp | 1 |
| [AA AutomationBench → CWE-Bench-AA](https://benchgap.net/calibration/aa-automation-bench/current/to/aa-cwe-bench/current.md) | Hill | 11 | 9.0 pp | 3 |
| [MathVision → VideoMMMU](https://benchgap.net/calibration/mathvision/current/to/videommmu/current.md) | inverse Michaelis–Menten | 10 | 9.0 pp | 0 |
| [MLVU (M-Avg) → ERQA](https://benchgap.net/calibration/mlvuavg/current/to/erqa/current.md) | Hill | 8 | 9.0 pp | 0 |
| [AA Agentic Index → Hermes Index](https://benchgap.net/calibration/aa-agentic-index/current/to/hermesindex/current.md) | Michaelis–Menten + offset | 5 | 9.0 pp | 0 |
| [GDP.pdf → DeepsecBench-AA](https://benchgap.net/calibration/aa-gdp-pdf/current/to/aa-deepsec-bench/current.md) | Michaelis–Menten + offset | 11 | 9.1 pp | 0 |
| [SkillsBench official leaderboard (BenchFlow) → EBR-bench](https://benchgap.net/calibration/skillsbench/current/to/ebrbench/current.md) | linear | 10 | 9.1 pp | 6 |
| [VITA-Bench → DeepPlanning](https://benchgap.net/calibration/vitabench/current/to/deepplanning/current.md) | offset logistic | 5 | 9.1 pp | 8 |
| [Decision Index Retrieval → Decision Index Tools](https://benchgap.net/calibration/jevdecisionindex03retrieval/current/to/jevdecisionindex03tools/current.md) | linear | 110 | 9.1 pp | 0 |
| [Claw-Eval → Gert Labs](https://benchgap.net/calibration/claw-eval/current/to/gertlabs/current.md) | Michaelis–Menten + offset | 22 | 9.1 pp | 6 |
| [MCP Atlas → SkillsBench](https://benchgap.net/calibration/mcpatlas/current/to/skills-bench/current.md) | Michaelis–Menten | 5 | 9.1 pp | 13 |
| [HLE → AA-Omniscience Accuracy](https://benchgap.net/calibration/hle/current/to/omniscienceaccuracy/current.md) | offset logistic | 54 | 9.1 pp | 4 |
| [Bug Hunt Bench → SWE-bench Pro](https://benchgap.net/calibration/bug-hunt-bench/current/to/swe-bench-pro/current.md) | linear | 8 | 9.1 pp | 0 |
| [MMMU-Pro → BabyVision](https://benchgap.net/calibration/mmmu-pro/current/to/babyvision/current.md) | Hill | 6 | 9.1 pp | 26 |
| [RealWorldQA → MathVision](https://benchgap.net/calibration/realworldqa/current/to/mathvision/current.md) | Hill | 11 | 9.1 pp | 1 |
| [AA Coding Index → ScarfBench](https://benchgap.net/calibration/aa-coding-index/current/to/scarfbench/current.md) | offset logistic | 6 | 9.1 pp | 69 |
| [MMLU-Pro → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/mmlu-pro/current/to/artificialanalysis/current.md) | Hill | 29 | 9.1 pp | 0 |
| [Terminal-Bench 2.1 → Terminal-Bench 3.0](https://benchgap.net/calibration/terminal-bench%23terminal-bench-2-1/current/to/terminal-bench%23terminal-bench-3-0/current.md) | inverse Michaelis–Menten | 11 | 9.1 pp | 43 |
| [AA Agentic Index → OSWorld-Verified](https://benchgap.net/calibration/aa-agentic-index/current/to/osworld-verified/current.md) | linear | 17 | 9.1 pp | 0 |
| [AA ITBench → Gert Labs](https://benchgap.net/calibration/aa-itbench/current/to/gertlabs/current.md) | Hill | 8 | 9.1 pp | 1 |
| [APEX-Agents-AA → MCP Atlas](https://benchgap.net/calibration/apexagentsaa/current/to/mcpatlas/current.md) | Michaelis–Menten + offset | 14 | 9.1 pp | 3 |
| [AA-HLE → AA-Omniscience Accuracy](https://benchgap.net/calibration/aa-hle/current/to/omniscienceaccuracy/current.md) | offset logistic | 159 | 9.1 pp | 0 |
| [τ³-bench results → AutomationBench](https://benchgap.net/calibration/tau3-bench/current/to/automationbench/current.md) | Michaelis–Menten | 5 | 9.1 pp | 26 |
| [HLE w/ tools → AA Agentic Index](https://benchgap.net/calibration/hlewithtools/current/to/aa-agentic-index/current.md) | inverse Michaelis–Menten | 10 | 9.2 pp | 2 |
| [GDPval-AA → BrowseComp](https://benchgap.net/calibration/gdpvalaanormalized/current/to/browsecomp/current.md) | Michaelis–Menten | 34 | 9.2 pp | 28 |
| [AA-HLE → AA-GPQA Diamond](https://benchgap.net/calibration/aa-hle/current/to/aa-gpqa-diamond/current.md) | Hill | 190 | 9.2 pp | 3 |
| [IFEval → IFBench](https://benchgap.net/calibration/ifeval/current/to/ifbench/current.md) | Hill | 22 | 9.2 pp | 22 |
| [BrowseComp → ExploitGym](https://benchgap.net/calibration/browsecomp/current/to/exploitgym/current.md) | linear | 7 | 9.2 pp | 25 |
| [SciCode → AA Coding Index](https://benchgap.net/calibration/scicode/current/to/aa-coding-index/current.md) | linear | 15 | 9.2 pp | 1 |
| [Vals MMLU-Pro → HLE](https://benchgap.net/calibration/vals-mmlu-pro/current/to/hle/current.md) | inverse Michaelis–Menten | 32 | 9.2 pp | 8 |
| [ERQA → SimpleVQA](https://benchgap.net/calibration/erqa/current/to/simplevqa/current.md) | Michaelis–Menten | 12 | 9.2 pp | 1 |
| [Mystery Game Puzzles → StealthMark](https://benchgap.net/calibration/mysterygamepuzzles/current/to/stealthmark/current.md) | linear | 29 | 9.2 pp | 19 |
| [Terminal-Bench Hard → Terminal-Bench 4.1](https://benchgap.net/calibration/terminal-bench-hard/current/to/terminal-bench41/current.md) | offset logistic | 5 | 9.2 pp | 15 |
| [AA Agentic Index → AA Cyber Index](https://benchgap.net/calibration/aa-agentic-index/current/to/aa-cyber-index/current.md) | Michaelis–Menten | 8 | 9.2 pp | 61 |
| [Mystery Game Puzzles → Chess Puzzles](https://benchgap.net/calibration/mysterygamepuzzles/current/to/chesspuzzles/current.md) | Michaelis–Menten | 63 | 9.3 pp | 0 |
| [Claw-Eval → JobBench](https://benchgap.net/calibration/claw-eval/current/to/jobbench/current.md) | linear | 7 | 9.3 pp | 10 |
| [AA-Omniscience Accuracy → SuperGPQA](https://benchgap.net/calibration/omniscienceaccuracy/current/to/supergpqa/current.md) | Hill | 12 | 9.3 pp | 85 |
| [StealthMark → SEC-Bench Pro](https://benchgap.net/calibration/stealthmark/current/to/secbenchpro/current.md) | linear | 11 | 9.3 pp | 15 |
| [CyberGym → ExploitGym](https://benchgap.net/calibration/cybergym/current/to/exploitgym/current.md) | offset logistic | 11 | 9.3 pp | 7 |
| [AA-SciCode → FrontierCode 2.0](https://benchgap.net/calibration/aa-sci-code/current/to/frontiercode-2/current.md) | offset logistic | 21 | 9.3 pp | 31 |
| [Furniture Assembly → EBR-bench](https://benchgap.net/calibration/furnitureassembly/current/to/ebrbench/current.md) | linear | 27 | 9.3 pp | 12 |
| [MathVision → ScreenSpot Pro](https://benchgap.net/calibration/mathvision/current/to/screenspot-pro/current.md) | offset logistic | 9 | 9.3 pp | 1 |
| [BlenderWorld (text-only) → Chess Puzzles](https://benchgap.net/calibration/blenderworld-text-only/current/to/chesspuzzles/current.md) | Hill | 5 | 9.3 pp | 0 |
| [Mystery Game Puzzles → Furniture Assembly](https://benchgap.net/calibration/mysterygamepuzzles/current/to/furnitureassembly/current.md) | Hill | 39 | 9.3 pp | 28 |
| [HLE w/ tools → Hermes Index](https://benchgap.net/calibration/hlewithtools/current/to/hermesindex/current.md) | linear | 8 | 9.3 pp | 9 |
| [OSWorld-Verified → Toolathlon](https://benchgap.net/calibration/osworld-verified/current/to/toolathlon/current.md) | Hill | 11 | 9.3 pp | 4 |
| [Gert Labs → CyberGym](https://benchgap.net/calibration/gertlabs/current/to/cybergym/current.md) | inverse Michaelis–Menten | 8 | 9.3 pp | 17 |
| [CritPt → AA-GPQA Diamond](https://benchgap.net/calibration/critpt/current/to/aa-gpqa-diamond/current.md) | Hill | 176 | 9.3 pp | 2 |
| [GPQA → ARC-AGI-3](https://benchgap.net/calibration/gpqa/current/to/arcagi3/current.md) | offset logistic | 12 | 9.3 pp | 4 |
| [Mystery Game Puzzles → SpireClimb](https://benchgap.net/calibration/mysterygamepuzzles/current/to/spireclimb/current.md) | Hill | 12 | 9.3 pp | 22 |
| [LiveCodeBench → SWE-bench Verified](https://benchgap.net/calibration/livecodebench/current/to/swe-bench-verified/current.md) | Michaelis–Menten | 5 | 9.3 pp | 9 |
| [Mystery Game Puzzles → SEC-Bench Pro](https://benchgap.net/calibration/mysterygamepuzzles/current/to/secbenchpro/current.md) | linear | 11 | 9.3 pp | 22 |
| [Video-MME (w/o subtitle) → ERQA](https://benchgap.net/calibration/videommenosub/current/to/erqa/current.md) | Michaelis–Menten | 5 | 9.3 pp | 0 |
| [Agents' Last Exam → DeepsecBench-AA](https://benchgap.net/calibration/agentslastexam/current/to/aa-deepsec-bench/current.md) | Michaelis–Menten | 9 | 9.4 pp | 1 |
| [Decision Index Arts → Decision Index Tools](https://benchgap.net/calibration/jevdecisionindex03arts/current/to/jevdecisionindex03tools/current.md) | Michaelis–Menten + offset | 110 | 9.4 pp | 0 |
| [Decision Index Language → When2Call (Decision Index)](https://benchgap.net/calibration/jevdecisionindex03language/current/to/aa-decision-when2-call/current.md) | Michaelis–Menten | 11 | 9.4 pp | 1 |
| [Artificial Analysis Intelligence Index → AA-HLE](https://benchgap.net/calibration/artificialanalysis/current/to/aa-hle/current.md) | Michaelis–Menten + offset | 167 | 9.4 pp | 4 |
| [MCP Atlas → OSWorld-Verified](https://benchgap.net/calibration/mcpatlas/current/to/osworld-verified/current.md) | Hill | 15 | 9.4 pp | 2 |
| [ARC-AGI-3 Provider Adapter → ARC-AGI-2](https://benchgap.net/calibration/arc-agi-3-provider-adapter/current/to/arc-agi-2/current.md) | Michaelis–Menten + offset | 12 | 9.4 pp | 1 |
| [GDPval-AA → APEX-Agents](https://benchgap.net/calibration/gdpvalaanormalized/current/to/apexagents/current.md) | inverse Michaelis–Menten | 7 | 9.4 pp | 94 |
| [AutomationBench → AA Cyber Index](https://benchgap.net/calibration/automationbench/current/to/aa-cyber-index/current.md) | offset logistic | 10 | 9.4 pp | 16 |
| [FrontierPhysics → Chess Puzzles](https://benchgap.net/calibration/frontierphysics/current/to/chesspuzzles/current.md) | linear | 24 | 9.4 pp | 1 |
| [CyberGym-E2E-AA → AutomationBench](https://benchgap.net/calibration/aa-cyber-gym-e2e/current/to/automationbench/current.md) | offset logistic | 10 | 9.5 pp | 2 |
| [Vals SWE-bench → FrontierSWE v2](https://benchgap.net/calibration/valsswebench/current/to/frontierswev2/current.md) | inverse Michaelis–Menten | 12 | 9.5 pp | 21 |
| [Decision Index Retrieval → Decision Index Language](https://benchgap.net/calibration/jevdecisionindex03retrieval/current/to/jevdecisionindex03language/current.md) | linear | 110 | 9.5 pp | 0 |
| [CWE-Bench-AA → DeepsecBench-AA](https://benchgap.net/calibration/aa-cwe-bench/current/to/aa-deepsec-bench/current.md) | offset logistic | 19 | 9.5 pp | 0 |
| [Decision Index Knowledge → When2Call (Decision Index)](https://benchgap.net/calibration/jevdecisionindex03knowledge/current/to/aa-decision-when2-call/current.md) | linear | 11 | 9.5 pp | 2 |
| [CharXiv w/o tools → MathVision](https://benchgap.net/calibration/charxivnotools/current/to/mathvision/current.md) | Hill | 9 | 9.5 pp | 0 |
| [ScarfBench → AA Coding Index](https://benchgap.net/calibration/scarfbench/current/to/aa-coding-index/current.md) | Hill | 6 | 9.5 pp | 0 |
| [BinaryAudit → SkillsBench official leaderboard (BenchFlow)](https://benchgap.net/calibration/binaryaudit/current/to/skillsbench/current.md) | Michaelis–Menten | 7 | 9.5 pp | 6 |
| [Decision Index Language → Decision Index Retrieval](https://benchgap.net/calibration/jevdecisionindex03language/current/to/jevdecisionindex03retrieval/current.md) | linear | 110 | 9.5 pp | 0 |
| [GDPval-AA → APEX-Agents-AA](https://benchgap.net/calibration/gdpvalaanormalized/current/to/apexagentsaa/current.md) | linear | 24 | 9.5 pp | 55 |
| [SimpleQA → GPQA](https://benchgap.net/calibration/simpleqa/current/to/gpqa/current.md) | Michaelis–Menten | 15 | 9.6 pp | 0 |
| [OCRBench V2 → VideoMMMU](https://benchgap.net/calibration/ocrbenchv2/current/to/videommmu/current.md) | Hill | 6 | 9.6 pp | 1 |
| [GDPval-AA → DeepSearchQA](https://benchgap.net/calibration/gdpvalaanormalized/current/to/deepsearchqa/current.md) | offset logistic | 15 | 9.6 pp | 39 |
| [Toolathlon → GDPval-AA](https://benchgap.net/calibration/toolathlon/current/to/gdpvalaanormalized/current.md) | Michaelis–Menten + offset | 19 | 9.6 pp | 0 |
| [MMMU-Pro → ScreenSpot Pro](https://benchgap.net/calibration/mmmu-pro/current/to/screenspot-pro/current.md) | Michaelis–Menten + offset | 15 | 9.6 pp | 18 |
| [FACTS Multimodal → MaxxBench](https://benchgap.net/calibration/facts-multimodal/current/to/maxxbench/current.md) | Hill | 7 | 9.6 pp | 0 |
| [ExploitGym → Agents' Last Exam](https://benchgap.net/calibration/exploitgym/current/to/agentslastexam/current.md) | offset logistic | 8 | 9.6 pp | 8 |
| [VideoMMMU → ERQA](https://benchgap.net/calibration/videommmu/current/to/erqa/current.md) | Michaelis–Menten + offset | 8 | 9.6 pp | 0 |
| [MirrorCode → MaxxBench](https://benchgap.net/calibration/mirrorcode/current/to/maxxbench/current.md) | linear | 10 | 9.7 pp | 0 |
| [AA AutomationBench → Hermes Index](https://benchgap.net/calibration/aa-automation-bench/current/to/hermesindex/current.md) | Hill | 6 | 9.7 pp | 0 |
| [LiveCodeBench v6 → SWE-bench Verified](https://benchgap.net/calibration/livecodebench-v6/current/to/swe-bench-verified/current.md) | offset logistic | 32 | 9.7 pp | 10 |
| [GPQA Diamond → ARC-AGI-2](https://benchgap.net/calibration/gpqa-diamond/current/to/arc-agi-2/current.md) | linear | 19 | 9.7 pp | 47 |
| [MCP Atlas → HLE w/ tools](https://benchgap.net/calibration/mcpatlas/current/to/hlewithtools/current.md) | Michaelis–Menten | 19 | 9.7 pp | 1 |
| [Chess Puzzles → StealthMark](https://benchgap.net/calibration/chesspuzzles/current/to/stealthmark/current.md) | linear | 23 | 9.7 pp | 17 |
| [GPQA Diamond → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/gpqa-diamond/current/to/artificialanalysis/current.md) | inverse Michaelis–Menten | 42 | 9.8 pp | 3 |
| [SEC-Bench Pro → Chess Puzzles](https://benchgap.net/calibration/secbenchpro/current/to/chesspuzzles/current.md) | Michaelis–Menten | 10 | 9.8 pp | 0 |
| [SkillsBench official leaderboard (BenchFlow) → Mystery Game Puzzles](https://benchgap.net/calibration/skillsbench/current/to/mysterygamepuzzles/current.md) | Hill | 14 | 9.8 pp | 5 |
| [EBR-bench → ExploitBench Internal Port](https://benchgap.net/calibration/ebrbench/current/to/exploitbenchinternalport/current.md) | linear | 6 | 9.8 pp | 0 |
| [BrowseComp → AA Harvey LAB v1.0](https://benchgap.net/calibration/browsecomp/current/to/aa-harvey-lab/current.md) | Michaelis–Menten + offset | 5 | 9.8 pp | 24 |
| [MCP Atlas → CyberGym](https://benchgap.net/calibration/mcpatlas/current/to/cybergym/current.md) | offset logistic | 12 | 9.8 pp | 2 |
| [AA AutomationBench → ExploitGym](https://benchgap.net/calibration/aa-automation-bench/current/to/exploitgym/current.md) | Hill | 7 | 9.8 pp | 1 |
| [Cookie-Bench → NYT Connections](https://benchgap.net/calibration/cookie-bench/current/to/nyt-connections/current.md) | Michaelis–Menten + offset | 5 | 9.8 pp | 0 |
| [AA-HLE → ARC-AGI-3](https://benchgap.net/calibration/aa-hle/current/to/arcagi3/current.md) | offset logistic | 22 | 9.8 pp | 116 |
| [CWE-bench → DeepsecBench-AA](https://benchgap.net/calibration/cwe-bench-v2/current/to/aa-deepsec-bench/current.md) | Michaelis–Menten | 9 | 9.8 pp | 0 |
| [HealthBench (raw) → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/healthbench/current/to/artificialanalysis/current.md) | Michaelis–Menten + offset | 8 | 9.8 pp | 0 |
| [Vals MMLU-Pro → AA-HLE](https://benchgap.net/calibration/vals-mmlu-pro/current/to/aa-hle/current.md) | offset logistic | 74 | 9.8 pp | 0 |
| [MCP Atlas → VITA-Bench](https://benchgap.net/calibration/mcpatlas/current/to/vitabench/current.md) | Michaelis–Menten | 7 | 9.9 pp | 13 |
| [SEC-Bench Pro → EBR-bench](https://benchgap.net/calibration/secbenchpro/current/to/ebrbench/current.md) | Hill | 9 | 9.9 pp | 2 |
| [AA-HLE → GPQA](https://benchgap.net/calibration/aa-hle/current/to/gpqa/current.md) | offset logistic | 91 | 9.9 pp | 0 |
| [ARC-AGI-3 → ARC-AGI-2](https://benchgap.net/calibration/arcagi3/current/to/arc-agi-2/current.md) | Hill | 24 | 9.9 pp | 0 |
| [Mystery Game Puzzles → CompileBench](https://benchgap.net/calibration/mysterygamepuzzles/current/to/compilebench/current.md) | linear | 5 | 10.0 pp | 51 |
| [ScarfBench → Bug Hunt Bench](https://benchgap.net/calibration/scarfbench/current/to/bug-hunt-bench/current.md) | Hill | 9 | 10.0 pp | 0 |
| [ExploitGym → CyberGym](https://benchgap.net/calibration/exploitgym/current/to/cybergym/current.md) | Michaelis–Menten | 11 | 10.0 pp | 4 |
| [ProgramBench → AA Coding Index](https://benchgap.net/calibration/programbench/current/to/aa-coding-index/current.md) | offset logistic | 6 | 10.0 pp | 0 |
| [ApprenticeBench → JobBench](https://benchgap.net/calibration/apprenticebench/current/to/jobbench/current.md) | linear | 7 | 10.0 pp | 0 |
| [SWE-bench Verified → Vals LiveCodeBench](https://benchgap.net/calibration/swe-bench-verified/current/to/valslivecodebench/current.md) | Hill | 32 | 10.0 pp | 30 |
| [AA Coding Index → Vibe Code Bench](https://benchgap.net/calibration/aa-coding-index/current/to/vibecodebench/current.md) | inverse Michaelis–Menten | 16 | 10.0 pp | 52 |
| [AA-Omniscience Accuracy → HLE](https://benchgap.net/calibration/omniscienceaccuracy/current/to/hle/current.md) | linear | 54 | 10.0 pp | 4 |
| [CritPt → AA-Omniscience Accuracy](https://benchgap.net/calibration/critpt/current/to/omniscienceaccuracy/current.md) | linear | 159 | 10.0 pp | 1 |
| [BioMysteryBench (human-difficult) → AA-HLE](https://benchgap.net/calibration/biomysterybenchhumandifficult/current/to/aa-hle/current.md) | offset logistic | 6 | 10.1 pp | 0 |
| [RealWorldQA → V*](https://benchgap.net/calibration/realworldqa/current/to/vstar/current.md) | Hill | 6 | 10.1 pp | 2 |
| [GDPval-AA → CWE-Bench-AA](https://benchgap.net/calibration/gdpvalaanormalized/current/to/aa-cwe-bench/current.md) | Michaelis–Menten + offset | 18 | 10.1 pp | 71 |
| [ParseBench-1 → ScreenSpot Pro](https://benchgap.net/calibration/parse-bench-llama-index/current/to/screenspot-pro/current.md) | Michaelis–Menten | 6 | 10.1 pp | 16 |
| [DeepSearchQA → ApprenticeBench](https://benchgap.net/calibration/deepsearchqa/current/to/apprenticebench/current.md) | Hill | 5 | 10.1 pp | 8 |
| [AA-GPQA Diamond → HLE](https://benchgap.net/calibration/aa-gpqa-diamond/current/to/hle/current.md) | inverse Michaelis–Menten | 51 | 10.1 pp | 3 |
| [EBR-bench → Chess Puzzles](https://benchgap.net/calibration/ebrbench/current/to/chesspuzzles/current.md) | linear | 30 | 10.1 pp | 0 |
| [FrontierCode 1.1 Main → FrontierCode 2.0](https://benchgap.net/calibration/frontiercode/current/to/frontiercode-2/current.md) | linear | 10 | 10.1 pp | 0 |
| [SciCode → SWE-bench Pro](https://benchgap.net/calibration/scicode/current/to/swe-bench-pro/current.md) | Hill | 20 | 10.1 pp | 1 |
| [DAYJOB Healthcare → ExploitBench Internal Port](https://benchgap.net/calibration/dayjob-healthcare/current/to/exploitbenchinternalport/current.md) | Michaelis–Menten | 5 | 10.1 pp | 0 |
| [HealthBench Professional (length-adjusted, Artificial Analysis) → AA-HLE](https://benchgap.net/calibration/aa-health-bench-professional/current/to/aa-hle/current.md) | inverse Michaelis–Menten | 6 | 10.1 pp | 0 |
| [OCRBench V2 → MMMU](https://benchgap.net/calibration/ocrbenchv2/current/to/mmmu/current.md) | Michaelis–Menten + offset | 5 | 10.1 pp | 0 |
| [Artificial Analysis Intelligence Index → SuperGPQA](https://benchgap.net/calibration/artificialanalysis/current/to/supergpqa/current.md) | offset logistic | 12 | 10.1 pp | 12 |
| [MCP Atlas → GDPval-AA](https://benchgap.net/calibration/mcpatlas/current/to/gdpvalaanormalized/current.md) | inverse Michaelis–Menten | 32 | 10.2 pp | 1 |
| [ResearchClawBench → DeepSearchQA](https://benchgap.net/calibration/researchclawbench/current/to/deepsearchqa/current.md) | Michaelis–Menten | 6 | 10.2 pp | 3 |
| [MaxxBench → Chess Puzzles](https://benchgap.net/calibration/maxxbench/current/to/chesspuzzles/current.md) | Hill | 14 | 10.2 pp | 0 |
| [ResearchClawBench → APEX-Agents-AA](https://benchgap.net/calibration/researchclawbench/current/to/apexagentsaa/current.md) | Michaelis–Menten | 9 | 10.2 pp | 3 |
| [AA-AnalystAgent → ExploitGym](https://benchgap.net/calibration/aa-analyst-agent/current/to/exploitgym/current.md) | linear | 10 | 10.2 pp | 0 |
| [AA-Omniscience Accuracy → GPQA](https://benchgap.net/calibration/omniscienceaccuracy/current/to/gpqa/current.md) | Michaelis–Menten | 72 | 10.2 pp | 0 |
| [Toolathlon → OSWorld-Verified](https://benchgap.net/calibration/toolathlon/current/to/osworld-verified/current.md) | Michaelis–Menten | 11 | 10.2 pp | 0 |
| [Toolathlon-Verified → Hermes Index](https://benchgap.net/calibration/toolathlonverified/current/to/hermesindex/current.md) | Hill | 8 | 10.2 pp | 1 |
| [DeepSWE → FrontierCode 2.0](https://benchgap.net/calibration/deepswe/current/to/frontiercode-2/current.md) | offset logistic | 22 | 10.2 pp | 11 |
| [Artificial Analysis Intelligence Index → BioBasicsBench](https://benchgap.net/calibration/artificialanalysis/current/to/biobasicsbench/current.md) | Michaelis–Menten | 7 | 10.2 pp | 39 |
| [ApprenticeBench → DeepsecBench-AA](https://benchgap.net/calibration/apprenticebench/current/to/aa-deepsec-bench/current.md) | Hill | 6 | 10.2 pp | 1 |
| [MaxxBench → NYT Connections](https://benchgap.net/calibration/maxxbench/current/to/nyt-connections/current.md) | linear | 12 | 10.2 pp | 1 |
| [MMMLU → AA-HLE](https://benchgap.net/calibration/mmmlu/current/to/aa-hle/current.md) | inverse Michaelis–Menten | 9 | 10.3 pp | 0 |
| [SimpleVQA → VideoMMMU](https://benchgap.net/calibration/simplevqa/current/to/videommmu/current.md) | Hill | 11 | 10.3 pp | 2 |
| [ExploitBench → NYT Connections](https://benchgap.net/calibration/exploitbench/current/to/nyt-connections/current.md) | offset logistic | 9 | 10.3 pp | 4 |
| [AA-Omniscience Accuracy → AA-GPQA Diamond](https://benchgap.net/calibration/omniscienceaccuracy/current/to/aa-gpqa-diamond/current.md) | Michaelis–Menten | 136 | 10.3 pp | 3 |
| [Gert Labs → JobBench](https://benchgap.net/calibration/gertlabs/current/to/jobbench/current.md) | inverse Michaelis–Menten | 15 | 10.3 pp | 9 |
| [SpireClimb → ExploitBench](https://benchgap.net/calibration/spireclimb/current/to/exploitbench/current.md) | linear | 5 | 10.3 pp | 0 |
| [BrowseComp → MCP Atlas](https://benchgap.net/calibration/browsecomp/current/to/mcpatlas/current.md) | offset logistic | 35 | 10.3 pp | 13 |
| [Chess Puzzles → MaxxBench](https://benchgap.net/calibration/chesspuzzles/current/to/maxxbench/current.md) | linear | 14 | 10.3 pp | 17 |
| [Chess Puzzles → SpireClimb](https://benchgap.net/calibration/chesspuzzles/current/to/spireclimb/current.md) | linear | 11 | 10.3 pp | 13 |
| [CWE-Bench-AA → AA Cyber Index](https://benchgap.net/calibration/aa-cwe-bench/current/to/aa-cyber-index/current.md) | Hill | 21 | 10.3 pp | 0 |
| [Chess Puzzles → Mystery Game Puzzles](https://benchgap.net/calibration/chesspuzzles/current/to/mysterygamepuzzles/current.md) | inverse Michaelis–Menten | 63 | 10.3 pp | 14 |
| [MazeBench → DeepsecBench-AA](https://benchgap.net/calibration/mazebench/current/to/aa-deepsec-bench/current.md) | linear | 17 | 10.3 pp | 3 |
| [MLVU (M-Avg) → SimpleVQA](https://benchgap.net/calibration/mlvuavg/current/to/simplevqa/current.md) | inverse Michaelis–Menten | 10 | 10.3 pp | 0 |
| [FrontierPhysics → ExploitBench](https://benchgap.net/calibration/frontierphysics/current/to/exploitbench/current.md) | Hill | 12 | 10.4 pp | 1 |
| [Toolathlon → AA Agentic Index](https://benchgap.net/calibration/toolathlon/current/to/aa-agentic-index/current.md) | offset logistic | 15 | 10.4 pp | 0 |
| [DeepSWE → Bug Hunt Bench](https://benchgap.net/calibration/deepswe/current/to/bug-hunt-bench/current.md) | Hill | 24 | 10.4 pp | 17 |
| [CritPt → HLE](https://benchgap.net/calibration/critpt/current/to/hle/current.md) | Hill | 61 | 10.4 pp | 1 |
| [DynaMath → MMMU](https://benchgap.net/calibration/dynamath/current/to/mmmu/current.md) | Hill | 7 | 10.4 pp | 0 |
| [OSWorld 2.0 → ExploitGym](https://benchgap.net/calibration/osworld2/current/to/exploitgym/current.md) | Michaelis–Menten | 10 | 10.4 pp | 2 |
| [OSWorld 2.0 → MCP Atlas](https://benchgap.net/calibration/osworld2/current/to/mcpatlas/current.md) | Michaelis–Menten | 8 | 10.4 pp | 2 |
| [APEX-Agents-AA → Gert Labs](https://benchgap.net/calibration/apexagentsaa/current/to/gertlabs/current.md) | linear | 13 | 10.4 pp | 3 |
| [PostTrainBench v1.1 → FrontierSWE v2](https://benchgap.net/calibration/posttrainbench%23posttrainbench-v1-1/current/to/frontierswev2/current.md) | Hill | 7 | 10.4 pp | 1 |
| [Mystery Game Puzzles → EBR-bench](https://benchgap.net/calibration/mysterygamepuzzles/current/to/ebrbench/current.md) | linear | 32 | 10.4 pp | 28 |
| [Vals GPQA Diamond → HLE](https://benchgap.net/calibration/vals-gpqa-diamond/current/to/hle/current.md) | inverse Michaelis–Menten | 32 | 10.4 pp | 0 |
| [CyberGym → AutomationBench](https://benchgap.net/calibration/cybergym/current/to/automationbench/current.md) | Hill | 11 | 10.4 pp | 10 |
| [DynaMath → MathVision](https://benchgap.net/calibration/dynamath/current/to/mathvision/current.md) | Hill | 6 | 10.5 pp | 0 |
| [Agents' Last Exam → ExploitGym](https://benchgap.net/calibration/agentslastexam/current/to/exploitgym/current.md) | Michaelis–Menten + offset | 8 | 10.5 pp | 1 |
| [Vibe Code Bench → PostTrainBench v1.1](https://benchgap.net/calibration/vibecodebench/current/to/posttrainbench%23posttrainbench-v1-1/current.md) | Hill | 5 | 10.5 pp | 9 |
| [MMLU-Redux → AA-HLE](https://benchgap.net/calibration/mmluredux/current/to/aa-hle/current.md) | Hill | 9 | 10.5 pp | 0 |
| [ApprenticeBench → ExploitGym](https://benchgap.net/calibration/apprenticebench/current/to/exploitgym/current.md) | Michaelis–Menten | 6 | 10.5 pp | 0 |
| [AA-SciCode → Bug Hunt Bench](https://benchgap.net/calibration/aa-sci-code/current/to/bug-hunt-bench/current.md) | inverse Michaelis–Menten | 23 | 10.5 pp | 38 |
| [HealthBench Hard → AA-HLE](https://benchgap.net/calibration/healthbench-hard/current/to/aa-hle/current.md) | Michaelis–Menten | 12 | 10.5 pp | 0 |
| [CWE-Bench-AA → BrowseComp](https://benchgap.net/calibration/aa-cwe-bench/current/to/browsecomp/current.md) | Michaelis–Menten | 8 | 10.6 pp | 0 |
| [ProgramBench → Vals SWE-bench](https://benchgap.net/calibration/programbench/current/to/valsswebench/current.md) | offset logistic | 5 | 10.6 pp | 1 |
| [We-Math → MathVision](https://benchgap.net/calibration/wemath/current/to/mathvision/current.md) | Michaelis–Menten + offset | 6 | 10.6 pp | 0 |
| [JevBench 1.5 → JevBench 1.4](https://benchgap.net/calibration/jev-bench15/current/to/jev-bench14/current.md) | offset logistic | 86 | 10.6 pp | 18 |
| [CWE-bench → ExploitGym](https://benchgap.net/calibration/cwe-bench-v2/current/to/exploitgym/current.md) | linear | 7 | 10.6 pp | 0 |
| [GDPval-AA → ExploitGym](https://benchgap.net/calibration/gdpvalaanormalized/current/to/exploitgym/current.md) | Hill | 17 | 10.6 pp | 37 |
| [Toolathlon-Verified → MKA](https://benchgap.net/calibration/toolathlonverified/current/to/mka/current.md) | Hill | 5 | 10.6 pp | 2 |
| [MazeBench → CWE-Bench-AA](https://benchgap.net/calibration/mazebench/current/to/aa-cwe-bench/current.md) | Michaelis–Menten + offset | 17 | 10.6 pp | 3 |
| [Claw-Eval → VITA-Bench](https://benchgap.net/calibration/claw-eval/current/to/vitabench/current.md) | inverse Michaelis–Menten | 7 | 10.7 pp | 17 |
| [NYT Connections → StealthMark](https://benchgap.net/calibration/nyt-connections/current/to/stealthmark/current.md) | offset logistic | 32 | 10.7 pp | 21 |
| [AA Agentic Index → APEX-Agents-AA](https://benchgap.net/calibration/aa-agentic-index/current/to/apexagentsaa/current.md) | Michaelis–Menten | 18 | 10.7 pp | 0 |
| [CritPt → BioBasicsBench](https://benchgap.net/calibration/critpt/current/to/biobasicsbench/current.md) | offset logistic | 7 | 10.7 pp | 36 |
| [Furniture Assembly → SEC-Bench Pro](https://benchgap.net/calibration/furnitureassembly/current/to/secbenchpro/current.md) | Michaelis–Menten | 11 | 10.7 pp | 1 |
| [cursorBench41 → Bug Hunt Bench](https://benchgap.net/calibration/cursor-bench41/current/to/bug-hunt-bench/current.md) | Hill | 12 | 10.7 pp | 0 |
| [AIME26 → IMOAnswerBench](https://benchgap.net/calibration/aime2026/current/to/imoanswerbench/current.md) | Michaelis–Menten | 5 | 10.7 pp | 11 |
| [Terminal-Bench 4.1 → Terminal-Bench-Science 0.1](https://benchgap.net/calibration/terminal-bench41/current/to/terminal-bench-science/current.md) | linear | 12 | 10.7 pp | 30 |
| [MKA → Agents' Last Exam](https://benchgap.net/calibration/mka/current/to/agentslastexam/current.md) | Michaelis–Menten | 8 | 10.7 pp | 10 |
| [HLE → Artificial Analysis Intelligence Index](https://benchgap.net/calibration/hle/current/to/artificialanalysis/current.md) | linear | 52 | 10.7 pp | 3 |
| [BrowseComp → JobBench](https://benchgap.net/calibration/browsecomp/current/to/jobbench/current.md) | linear | 12 | 10.7 pp | 7 |
| [DeepPlanning → MCP Atlas](https://benchgap.net/calibration/deepplanning/current/to/mcpatlas/current.md) | linear | 7 | 10.7 pp | 0 |
| [GDPval-AA → MCP Atlas](https://benchgap.net/calibration/gdpvalaanormalized/current/to/mcpatlas/current.md) | Hill | 32 | 10.7 pp | 33 |
| [ExploitBench Internal Port → NYT Connections](https://benchgap.net/calibration/exploitbenchinternalport/current/to/nyt-connections/current.md) | Michaelis–Menten + offset | 6 | 10.8 pp | 0 |
| [Furniture Assembly → Mystery Game Puzzles](https://benchgap.net/calibration/furnitureassembly/current/to/mysterygamepuzzles/current.md) | offset logistic | 39 | 10.8 pp | 0 |
| [Video-MME (with subtitle) → SimpleVQA](https://benchgap.net/calibration/videommewithsub/current/to/simplevqa/current.md) | inverse Michaelis–Menten | 10 | 10.8 pp | 0 |
| [StealthMark → Chess Puzzles](https://benchgap.net/calibration/stealthmark/current/to/chesspuzzles/current.md) | Michaelis–Menten | 23 | 10.8 pp | 11 |
| [FrontierPhysics → ExploitBench Internal Port](https://benchgap.net/calibration/frontierphysics/current/to/exploitbenchinternalport/current.md) | linear | 7 | 10.8 pp | 1 |
| [Hermes Bench → ExploitGym](https://benchgap.net/calibration/hermesbench/current/to/exploitgym/current.md) | Hill | 6 | 10.8 pp | 0 |
| [RealWorldQA → MMMU-Pro](https://benchgap.net/calibration/realworldqa/current/to/mmmu-pro/current.md) | Hill | 12 | 10.8 pp | 1 |
| [IMOAnswerBench → AIME26](https://benchgap.net/calibration/imoanswerbench/current/to/aime2026/current.md) | Michaelis–Menten | 5 | 10.8 pp | 1 |
| [AA-MMMU-Pro → ScreenSpot Pro](https://benchgap.net/calibration/aa-mmmu-pro/current/to/screenspot-pro/current.md) | Michaelis–Menten + offset | 13 | 10.8 pp | 46 |
| [GPQA → ARC-AGI-2](https://benchgap.net/calibration/gpqa/current/to/arc-agi-2/current.md) | linear | 21 | 10.9 pp | 45 |
| [OpenHarmony Bench v1.2 → DeepSWE](https://benchgap.net/calibration/openharmonybench/current/to/deepswe/current.md) | offset logistic | 5 | 10.9 pp | 0 |
| [Terminal-Bench 2.1 → Terminal-Bench 4.1](https://benchgap.net/calibration/terminal-bench%23terminal-bench-2-1/current/to/terminal-bench41/current.md) | linear | 23 | 10.9 pp | 49 |
| [ScreenSpot Pro → MathVision](https://benchgap.net/calibration/screenspot-pro/current/to/mathvision/current.md) | Michaelis–Menten | 9 | 10.9 pp | 4 |
| [We-Math → V*](https://benchgap.net/calibration/wemath/current/to/vstar/current.md) | Michaelis–Menten | 5 | 10.9 pp | 0 |
| [Agents' Last Exam → AutomationBench](https://benchgap.net/calibration/agentslastexam/current/to/automationbench/current.md) | offset logistic | 17 | 10.9 pp | 0 |
| [ERQA → Video-MME (w/o subtitle)](https://benchgap.net/calibration/erqa/current/to/videommenosub/current.md) | Michaelis–Menten | 5 | 10.9 pp | 1 |
| [DRACO → Hermes Index](https://benchgap.net/calibration/draco/current/to/hermesindex/current.md) | linear | 5 | 10.9 pp | 0 |
| [MazeBench → Agents' Last Exam](https://benchgap.net/calibration/mazebench/current/to/agentslastexam/current.md) | linear | 12 | 11.0 pp | 24 |
| [MaxxBench → EBR-bench](https://benchgap.net/calibration/maxxbench/current/to/ebrbench/current.md) | Hill | 5 | 11.0 pp | 0 |
| [τ²-bench results → GDPval-AA](https://benchgap.net/calibration/tau2-bench/current/to/gdpvalaanormalized/current.md) | Hill | 66 | 11.0 pp | 49 |
| [Furniture Assembly → Chess Puzzles](https://benchgap.net/calibration/furnitureassembly/current/to/chesspuzzles/current.md) | linear | 38 | 11.0 pp | 0 |
| [AIME 2025 → HMMT Feb 2026](https://benchgap.net/calibration/aime2025/current/to/hmmtfeb2026/current.md) | offset logistic | 5 | 11.0 pp | 35 |
| [HealthBench Professional → ARC-AGI-3 Provider Adapter](https://benchgap.net/calibration/healthbenchprofessional/current/to/arc-agi-3-provider-adapter/current.md) | offset logistic | 7 | 11.0 pp | 2 |
| [MCP Atlas → ApprenticeBench](https://benchgap.net/calibration/mcpatlas/current/to/apprenticebench/current.md) | Hill | 5 | 11.0 pp | 14 |
| [Hermes Index → AA AutomationBench](https://benchgap.net/calibration/hermesindex/current/to/aa-automation-bench/current.md) | offset logistic | 6 | 11.0 pp | 1 |
| [MKA → AA AutomationBench](https://benchgap.net/calibration/mka/current/to/aa-automation-bench/current.md) | Hill | 16 | 11.0 pp | 1 |
| [Bug Hunt Bench → OpenHarmony Bench v1.2](https://benchgap.net/calibration/bug-hunt-bench/current/to/openharmonybench/current.md) | offset logistic | 5 | 11.0 pp | 23 |
| [Harvey LAB-AA v1.1 → DeepsecBench-AA](https://benchgap.net/calibration/aa-harvey-lab-v11/current/to/aa-deepsec-bench/current.md) | offset logistic | 6 | 11.0 pp | 0 |
| [Terminal-Bench 2.1 → AA Terminal-Bench 4.0](https://benchgap.net/calibration/terminal-bench%23terminal-bench-2-1/current/to/aa-terminal-bench4/current.md) | offset logistic | 5 | 11.0 pp | 40 |
| [FACTS Multimodal → Chess Puzzles](https://benchgap.net/calibration/facts-multimodal/current/to/chesspuzzles/current.md) | inverse Michaelis–Menten | 20 | 11.0 pp | 0 |
| [BabaIsBench → Chess Puzzles](https://benchgap.net/calibration/babaisbench/current/to/chesspuzzles/current.md) | offset logistic | 26 | 11.1 pp | 1 |
| [FACTS Multimodal → StealthMark](https://benchgap.net/calibration/facts-multimodal/current/to/stealthmark/current.md) | offset logistic | 17 | 11.1 pp | 0 |
| [Bug Hunt Bench → FrontierSWE v2](https://benchgap.net/calibration/bug-hunt-bench/current/to/frontierswev2/current.md) | Michaelis–Menten | 20 | 11.1 pp | 2 |
| [BabaIsBench → MaxxBench](https://benchgap.net/calibration/babaisbench/current/to/maxxbench/current.md) | Michaelis–Menten + offset | 9 | 11.1 pp | 1 |
| [React Native Evals → AA Coding Index](https://benchgap.net/calibration/reactnativeevals/current/to/aa-coding-index/current.md) | Hill | 9 | 11.1 pp | 3 |
| [CyberGym → MCP Atlas](https://benchgap.net/calibration/cybergym/current/to/mcpatlas/current.md) | Hill | 12 | 11.1 pp | 4 |
| [OSWorld-Verified → JobBench](https://benchgap.net/calibration/osworld-verified/current/to/jobbench/current.md) | offset logistic | 15 | 11.1 pp | 4 |
| [WideResearch → CyberGym](https://benchgap.net/calibration/wideresearch/current/to/cybergym/current.md) | Hill | 5 | 11.1 pp | 0 |
| [Gert Labs → APEX-Agents-AA](https://benchgap.net/calibration/gertlabs/current/to/apexagentsaa/current.md) | linear | 13 | 11.2 pp | 9 |
| [HMMT Nov 2025 → FrontierMath v2 (Tiers 1-3)](https://benchgap.net/calibration/hmmtnov2025/current/to/frontiermathv2tiers13/current.md) | Hill | 6 | 11.2 pp | 1 |
| [PostTrainBench → FrontierSWE v2](https://benchgap.net/calibration/post-train-bench12/current/to/frontierswev2/current.md) | linear | 10 | 11.2 pp | 1 |
| [DeepSearchQA → Gert Labs](https://benchgap.net/calibration/deepsearchqa/current/to/gertlabs/current.md) | offset logistic | 8 | 11.2 pp | 1 |
| [SWE-bench Pro → DeepSWE](https://benchgap.net/calibration/swe-bench-pro/current/to/deepswe/current.md) | Michaelis–Menten + offset | 25 | 11.2 pp | 10 |
| [ERQA → V*](https://benchgap.net/calibration/erqa/current/to/vstar/current.md) | Michaelis–Menten | 6 | 11.2 pp | 4 |
| [AA-LCR → MRCRv2](https://benchgap.net/calibration/lcr/current/to/mrcrv2/current.md) | offset logistic | 9 | 11.3 pp | 202 |
| [OSWorld-Verified → CyberGym](https://benchgap.net/calibration/osworld-verified/current/to/cybergym/current.md) | Hill | 10 | 11.3 pp | 4 |
| [GDP.pdf → JobBench](https://benchgap.net/calibration/aa-gdp-pdf/current/to/jobbench/current.md) | Hill | 5 | 11.3 pp | 8 |
| [We-Math → MMMU-Pro](https://benchgap.net/calibration/wemath/current/to/mmmu-pro/current.md) | offset logistic | 6 | 11.3 pp | 0 |
| [OTelBench → SkillsBench official leaderboard (BenchFlow)](https://benchgap.net/calibration/otelbench/current/to/skillsbench/current.md) | linear | 5 | 11.3 pp | 1 |
| [StealthMark → EBR-bench](https://benchgap.net/calibration/stealthmark/current/to/ebrbench/current.md) | inverse Michaelis–Menten | 12 | 11.3 pp | 11 |
| [MazeBench → BrowseComp](https://benchgap.net/calibration/mazebench/current/to/browsecomp/current.md) | Michaelis–Menten + offset | 14 | 11.3 pp | 4 |
| [Hermes Bench → AA AutomationBench](https://benchgap.net/calibration/hermesbench/current/to/aa-automation-bench/current.md) | Hill | 6 | 11.4 pp | 0 |
| [AA-Omniscience Accuracy → AA-HLE](https://benchgap.net/calibration/omniscienceaccuracy/current/to/aa-hle/current.md) | linear | 159 | 11.4 pp | 0 |
| [DAYJOB → EBR-bench](https://benchgap.net/calibration/dayjob/current/to/ebrbench/current.md) | Hill | 11 | 11.4 pp | 0 |
| [MMLU-ProX → PolyMath](https://benchgap.net/calibration/mmluprox/current/to/polymath/current.md) | offset logistic | 5 | 11.4 pp | 7 |
| [JobBench → CyberGym](https://benchgap.net/calibration/jobbench/current/to/cybergym/current.md) | Michaelis–Menten | 12 | 11.4 pp | 3 |
| [AA Agentic Index → CWE-Bench-AA](https://benchgap.net/calibration/aa-agentic-index/current/to/aa-cwe-bench/current.md) | Michaelis–Menten | 8 | 11.4 pp | 0 |
| [ERQA → ScreenSpot Pro](https://benchgap.net/calibration/erqa/current/to/screenspot-pro/current.md) | Hill | 10 | 11.4 pp | 0 |
| [DeepsecBench-AA → Agents' Last Exam](https://benchgap.net/calibration/aa-deepsec-bench/current/to/agentslastexam/current.md) | Hill | 9 | 11.4 pp | 1 |
| [τ²-bench results → APEX-Agents-AA](https://benchgap.net/calibration/tau2-bench/current/to/apexagentsaa/current.md) | offset logistic | 21 | 11.4 pp | 35 |
| [Vals LiveCodeBench → AA Coding Index](https://benchgap.net/calibration/valslivecodebench/current/to/aa-coding-index/current.md) | linear | 49 | 11.5 pp | 5 |
| [SpireClimb → StealthMark](https://benchgap.net/calibration/spireclimb/current/to/stealthmark/current.md) | Michaelis–Menten + offset | 9 | 11.5 pp | 0 |
| [We-Math → MMMU](https://benchgap.net/calibration/wemath/current/to/mmmu/current.md) | Michaelis–Menten + offset | 6 | 11.5 pp | 0 |
| [CyberGym → DeepSearchQA](https://benchgap.net/calibration/cybergym/current/to/deepsearchqa/current.md) | offset logistic | 5 | 11.5 pp | 4 |
| [APEX-Agents-AA → GDPval-AA](https://benchgap.net/calibration/apexagentsaa/current/to/gdpvalaanormalized/current.md) | linear | 24 | 11.5 pp | 0 |
| [Hermes Index → AA Agentic Index](https://benchgap.net/calibration/hermesindex/current/to/aa-agentic-index/current.md) | Michaelis–Menten + offset | 5 | 11.5 pp | 0 |
| [CharXiv → BabyVision](https://benchgap.net/calibration/charxiv/current/to/babyvision/current.md) | offset logistic | 10 | 11.5 pp | 8 |
| [DeepSearchQA → APEX-Agents-AA](https://benchgap.net/calibration/deepsearchqa/current/to/apexagentsaa/current.md) | offset logistic | 7 | 11.5 pp | 1 |
| [BrowseComp → Agents' Last Exam](https://benchgap.net/calibration/browsecomp/current/to/agentslastexam/current.md) | Hill | 6 | 11.6 pp | 31 |
| [CharXiv w/o tools → ScreenSpot Pro](https://benchgap.net/calibration/charxivnotools/current/to/screenspot-pro/current.md) | inverse Michaelis–Menten | 5 | 11.6 pp | 0 |
| [HLE → ARC-AGI-1](https://benchgap.net/calibration/hle/current/to/arcagi1/current.md) | Michaelis–Menten | 16 | 11.6 pp | 3 |
| [SpireClimb → Furniture Assembly](https://benchgap.net/calibration/spireclimb/current/to/furnitureassembly/current.md) | Hill | 12 | 11.7 pp | 0 |
| [SimpleQA → AA-GPQA Diamond](https://benchgap.net/calibration/simpleqa/current/to/aa-gpqa-diamond/current.md) | linear | 9 | 11.7 pp | 0 |
| [FrontierPhysics → Mystery Game Puzzles](https://benchgap.net/calibration/frontierphysics/current/to/mysterygamepuzzles/current.md) | Hill | 25 | 11.7 pp | 1 |
| [MMLU-Pro → SimpleQA](https://benchgap.net/calibration/mmlu-pro/current/to/simpleqa/current.md) | linear | 8 | 11.7 pp | 14 |
| [HLE w/ tools → MCP Atlas](https://benchgap.net/calibration/hlewithtools/current/to/mcpatlas/current.md) | Michaelis–Menten | 19 | 11.7 pp | 0 |
| [HealthBench Professional → ARC-AGI-3](https://benchgap.net/calibration/healthbenchprofessional/current/to/arcagi3/current.md) | offset logistic | 12 | 11.8 pp | 1 |
| [MedXpertQA (MM) → CharXiv](https://benchgap.net/calibration/medxpertqamm/current/to/charxiv/current.md) | Michaelis–Menten | 6 | 11.8 pp | 0 |
| [PostTrainBench → Bug Hunt Bench](https://benchgap.net/calibration/post-train-bench12/current/to/bug-hunt-bench/current.md) | Hill | 13 | 11.8 pp | 0 |
| [Toolathlon → MCP Atlas](https://benchgap.net/calibration/toolathlon/current/to/mcpatlas/current.md) | inverse Michaelis–Menten | 17 | 11.8 pp | 0 |
| [BrowseComp → τ²-bench results](https://benchgap.net/calibration/browsecomp/current/to/tau2-bench/current.md) | Hill | 25 | 11.8 pp | 19 |
| [FrontierPhysics → EBR-bench](https://benchgap.net/calibration/frontierphysics/current/to/ebrbench/current.md) | Michaelis–Menten + offset | 15 | 11.8 pp | 1 |
| [DAYJOB → Chess Puzzles](https://benchgap.net/calibration/dayjob/current/to/chesspuzzles/current.md) | linear | 19 | 11.8 pp | 0 |
| [QwenClawBench → DeepPlanning](https://benchgap.net/calibration/qwenclawbench/current/to/deepplanning/current.md) | inverse Michaelis–Menten | 7 | 11.8 pp | 2 |
| [DAYJOB → Mystery Game Puzzles](https://benchgap.net/calibration/dayjob/current/to/mysterygamepuzzles/current.md) | Michaelis–Menten + offset | 19 | 11.9 pp | 0 |
| [Furniture Assembly → SpireClimb](https://benchgap.net/calibration/furnitureassembly/current/to/spireclimb/current.md) | linear | 12 | 11.9 pp | 0 |
| [BioMysteryBench (human-difficult) → CritPt](https://benchgap.net/calibration/biomysterybenchhumandifficult/current/to/critpt/current.md) | offset logistic | 6 | 11.9 pp | 0 |
| [CritPt → GPQA](https://benchgap.net/calibration/critpt/current/to/gpqa/current.md) | Hill | 84 | 11.9 pp | 1 |
| [ARC-AGI-1 → BioBasicsBench](https://benchgap.net/calibration/arcagi1/current/to/biobasicsbench/current.md) | offset logistic | 7 | 11.9 pp | 0 |
| [Artificial Analysis Intelligence Index → HLE](https://benchgap.net/calibration/artificialanalysis/current/to/hle/current.md) | linear | 52 | 12.0 pp | 1 |
| [SkillsBench official leaderboard (BenchFlow) → BinaryAudit](https://benchgap.net/calibration/skillsbench/current/to/binaryaudit/current.md) | inverse Michaelis–Menten | 7 | 12.0 pp | 13 |
| [AI2D_TEST → MMMU](https://benchgap.net/calibration/ai2dtest/current/to/mmmu/current.md) | Hill | 8 | 12.0 pp | 0 |
| [MMLU-Pro → Vals GPQA Diamond](https://benchgap.net/calibration/mmlu-pro/current/to/vals-gpqa-diamond/current.md) | inverse Michaelis–Menten | 10 | 12.1 pp | 1 |
| [MaxxBench → Furniture Assembly](https://benchgap.net/calibration/maxxbench/current/to/furnitureassembly/current.md) | inverse Michaelis–Menten | 10 | 12.1 pp | 1 |
| [NYT Connections → BinaryAudit](https://benchgap.net/calibration/nyt-connections/current/to/binaryaudit/current.md) | offset logistic | 11 | 12.1 pp | 61 |
| [BabaIsBench → NYT Connections](https://benchgap.net/calibration/babaisbench/current/to/nyt-connections/current.md) | linear | 26 | 12.1 pp | 1 |
| [SWE-bench Pro → Bug Hunt Bench](https://benchgap.net/calibration/swe-bench-pro/current/to/bug-hunt-bench/current.md) | Michaelis–Menten | 8 | 12.1 pp | 16 |
| [Artificial Analysis Intelligence Index → AA-GPQA Diamond](https://benchgap.net/calibration/artificialanalysis/current/to/aa-gpqa-diamond/current.md) | Hill | 147 | 12.1 pp | 1 |
| [AA Agentic Index → ExploitGym](https://benchgap.net/calibration/aa-agentic-index/current/to/exploitgym/current.md) | linear | 7 | 12.2 pp | 4 |
| [APEX-Agents-AA → AA Agentic Index](https://benchgap.net/calibration/apexagentsaa/current/to/aa-agentic-index/current.md) | Michaelis–Menten + offset | 18 | 12.2 pp | 2 |
| [DeepSearchQA → GDPval-AA](https://benchgap.net/calibration/deepsearchqa/current/to/gdpvalaanormalized/current.md) | linear | 15 | 12.2 pp | 1 |
| [QwenClawBench → MCP Atlas](https://benchgap.net/calibration/qwenclawbench/current/to/mcpatlas/current.md) | offset logistic | 8 | 12.3 pp | 1 |
| [ERQA → MMMU](https://benchgap.net/calibration/erqa/current/to/mmmu/current.md) | Michaelis–Menten | 6 | 12.3 pp | 0 |
| [Chess Puzzles → SEC-Bench Pro](https://benchgap.net/calibration/chesspuzzles/current/to/secbenchpro/current.md) | Michaelis–Menten | 10 | 12.3 pp | 18 |
| [Vals SWE-bench → ScarfBench](https://benchgap.net/calibration/valsswebench/current/to/scarfbench/current.md) | inverse Michaelis–Menten | 5 | 12.3 pp | 16 |
| [NYT Connections → Chess Puzzles](https://benchgap.net/calibration/nyt-connections/current/to/chesspuzzles/current.md) | Hill | 54 | 12.3 pp | 19 |
| [RealWorldQA → SimpleVQA](https://benchgap.net/calibration/realworldqa/current/to/simplevqa/current.md) | Hill | 11 | 12.3 pp | 2 |
| [ExploitBench → ExploitBench Internal Port](https://benchgap.net/calibration/exploitbench/current/to/exploitbenchinternalport/current.md) | Hill | 7 | 12.3 pp | 3 |
| [BFCL v4 → τ²-bench results](https://benchgap.net/calibration/bfcl-v4/current/to/tau2-bench/current.md) | Michaelis–Menten | 7 | 12.4 pp | 16 |
| [StealthMark → Furniture Assembly](https://benchgap.net/calibration/stealthmark/current/to/furnitureassembly/current.md) | Hill | 21 | 12.4 pp | 11 |
| [We-Math → SimpleVQA](https://benchgap.net/calibration/wemath/current/to/simplevqa/current.md) | Michaelis–Menten | 6 | 12.4 pp | 0 |
| [Toolathlon-Verified → APEX-Agents](https://benchgap.net/calibration/toolathlonverified/current/to/apexagents/current.md) | offset logistic | 5 | 12.4 pp | 4 |
| [BabaIsBench → Mystery Game Puzzles](https://benchgap.net/calibration/babaisbench/current/to/mysterygamepuzzles/current.md) | offset logistic | 26 | 12.4 pp | 1 |
| [MMLU-Pro → SuperGPQA](https://benchgap.net/calibration/mmlu-pro/current/to/supergpqa/current.md) | inverse Michaelis–Menten | 18 | 12.4 pp | 25 |
| [DeepSearchQA → AA Agentic Index](https://benchgap.net/calibration/deepsearchqa/current/to/aa-agentic-index/current.md) | Hill | 8 | 12.5 pp | 0 |
| [MazeBench → ApprenticeBench](https://benchgap.net/calibration/mazebench/current/to/apprenticebench/current.md) | linear | 16 | 12.5 pp | 22 |
| [AA Agentic Index → JobBench](https://benchgap.net/calibration/aa-agentic-index/current/to/jobbench/current.md) | offset logistic | 6 | 12.5 pp | 0 |
| [LongBench v2 → AA-LCR](https://benchgap.net/calibration/longbench-v2/current/to/lcr/current.md) | inverse Michaelis–Menten | 17 | 12.5 pp | 7 |
| [Hermes Bench → Agents' Last Exam](https://benchgap.net/calibration/hermesbench/current/to/agentslastexam/current.md) | inverse Michaelis–Menten | 7 | 12.6 pp | 0 |
| [Cookie-Bench → Mystery Game Puzzles](https://benchgap.net/calibration/cookie-bench/current/to/mysterygamepuzzles/current.md) | linear | 5 | 12.6 pp | 2 |
| [StealthMark → ExploitBench](https://benchgap.net/calibration/stealthmark/current/to/exploitbench/current.md) | inverse Michaelis–Menten | 11 | 12.6 pp | 9 |
| [Chess Puzzles → NYT Connections](https://benchgap.net/calibration/chesspuzzles/current/to/nyt-connections/current.md) | Michaelis–Menten | 54 | 12.6 pp | 24 |
| [MCP Atlas → APEX-Agents](https://benchgap.net/calibration/mcpatlas/current/to/apexagents/current.md) | linear | 6 | 12.6 pp | 10 |
| [AA Cyber Index → AA Agentic Index](https://benchgap.net/calibration/aa-cyber-index/current/to/aa-agentic-index/current.md) | linear | 8 | 12.7 pp | 2 |
| [Chess Puzzles → EBR-bench](https://benchgap.net/calibration/chesspuzzles/current/to/ebrbench/current.md) | inverse Michaelis–Menten | 30 | 12.7 pp | 15 |
| [SWE-bench Verified → DeepSWE](https://benchgap.net/calibration/swe-bench-verified/current/to/deepswe/current.md) | Hill | 12 | 12.7 pp | 27 |
| [STEMBench → AA-Omniscience Accuracy](https://benchgap.net/calibration/stembench/current/to/omniscienceaccuracy/current.md) | inverse Michaelis–Menten | 8 | 12.7 pp | 0 |
| [τ²-bench results → AA Agentic Index](https://benchgap.net/calibration/tau2-bench/current/to/aa-agentic-index/current.md) | linear | 40 | 12.7 pp | 51 |
| [SkillsBench official leaderboard (BenchFlow) → Chess Puzzles](https://benchgap.net/calibration/skillsbench/current/to/chesspuzzles/current.md) | linear | 17 | 12.8 pp | 0 |
| [AA Harvey LAB v1.0 → GDPval-AA](https://benchgap.net/calibration/aa-harvey-lab/current/to/gdpvalaanormalized/current.md) | offset logistic | 12 | 12.8 pp | 0 |
| [HLE w/ tools → AA Cyber Index](https://benchgap.net/calibration/hlewithtools/current/to/aa-cyber-index/current.md) | Michaelis–Menten + offset | 8 | 12.8 pp | 16 |
| [MazeBench → AA AutomationBench](https://benchgap.net/calibration/mazebench/current/to/aa-automation-bench/current.md) | Hill | 19 | 12.8 pp | 15 |
| [AA EnterpriseOps-Gym → JobBench](https://benchgap.net/calibration/aa-enterprise-ops-gym/current/to/jobbench/current.md) | Hill | 5 | 12.8 pp | 0 |
| [ProgramBench → SWE-Marathon](https://benchgap.net/calibration/programbench/current/to/swemarathon/current.md) | Michaelis–Menten | 7 | 12.9 pp | 7 |
| [ResearchClawBench → MCP Atlas](https://benchgap.net/calibration/researchclawbench/current/to/mcpatlas/current.md) | Hill | 13 | 12.9 pp | 3 |
| [SuperGPQA → GPQA Diamond](https://benchgap.net/calibration/supergpqa/current/to/gpqa-diamond/current.md) | Michaelis–Menten + offset | 7 | 12.9 pp | 2 |
| [BioBasicsBench → ARC-AGI-2](https://benchgap.net/calibration/biobasicsbench/current/to/arc-agi-2/current.md) | Michaelis–Menten + offset | 7 | 12.9 pp | 0 |
| [DAYJOB Healthcare → Chess Puzzles](https://benchgap.net/calibration/dayjob-healthcare/current/to/chesspuzzles/current.md) | linear | 19 | 12.9 pp | 0 |
| [GDP.pdf → Agents' Last Exam](https://benchgap.net/calibration/aa-gdp-pdf/current/to/agentslastexam/current.md) | Michaelis–Menten + offset | 5 | 12.9 pp | 1 |
| [CountBench → MMMU](https://benchgap.net/calibration/countbench/current/to/mmmu/current.md) | Hill | 10 | 13.0 pp | 0 |
| [HLE w/o tools → ARC-AGI-2](https://benchgap.net/calibration/hlenotools/current/to/arc-agi-2/current.md) | Hill | 18 | 13.0 pp | 8 |
| [τ²-bench results → BrowseComp](https://benchgap.net/calibration/tau2-bench/current/to/browsecomp/current.md) | Michaelis–Menten + offset | 25 | 13.0 pp | 50 |
| [MCP Atlas → DeepPlanning](https://benchgap.net/calibration/mcpatlas/current/to/deepplanning/current.md) | Michaelis–Menten + offset | 7 | 13.1 pp | 42 |
| [Vals LiveCodeBench → SWE-bench Verified](https://benchgap.net/calibration/valslivecodebench/current/to/swe-bench-verified/current.md) | inverse Michaelis–Menten | 32 | 13.1 pp | 5 |
| [StealthMark → Mystery Game Puzzles](https://benchgap.net/calibration/stealthmark/current/to/mysterygamepuzzles/current.md) | inverse Michaelis–Menten | 29 | 13.3 pp | 10 |
| [CWE-bench v1 → AA Agentic Index](https://benchgap.net/calibration/cwe-bench%23cwe-bench-v1/current/to/aa-agentic-index/current.md) | inverse Michaelis–Menten | 5 | 13.3 pp | 0 |
| [VITA-Bench → MCP Atlas](https://benchgap.net/calibration/vitabench/current/to/mcpatlas/current.md) | linear | 7 | 13.3 pp | 1 |
| [OSWorld 2.0 → AA Cyber Index](https://benchgap.net/calibration/osworld2/current/to/aa-cyber-index/current.md) | inverse Michaelis–Menten | 7 | 13.3 pp | 2 |
| [ZeroBench → BabyVision](https://benchgap.net/calibration/zerobench/current/to/babyvision/current.md) | Michaelis–Menten | 6 | 13.4 pp | 0 |
| [GPQA → SimpleQA](https://benchgap.net/calibration/gpqa/current/to/simpleqa/current.md) | offset logistic | 15 | 13.4 pp | 37 |
| [Chess Puzzles → ExploitBench](https://benchgap.net/calibration/chesspuzzles/current/to/exploitbench/current.md) | inverse Michaelis–Menten | 9 | 13.4 pp | 18 |
| [Claw-Eval → MCP Atlas](https://benchgap.net/calibration/claw-eval/current/to/mcpatlas/current.md) | linear | 19 | 13.5 pp | 7 |
| [DAYJOB Healthcare → SEC-Bench Pro](https://benchgap.net/calibration/dayjob-healthcare/current/to/secbenchpro/current.md) | linear | 8 | 13.5 pp | 0 |
| [BrowseComp → AA Cyber Index](https://benchgap.net/calibration/browsecomp/current/to/aa-cyber-index/current.md) | Michaelis–Menten + offset | 8 | 13.5 pp | 22 |
| [SEC-Bench Pro → Mystery Game Puzzles](https://benchgap.net/calibration/secbenchpro/current/to/mysterygamepuzzles/current.md) | inverse Michaelis–Menten | 11 | 13.6 pp | 0 |
| [BabaIsBench → EBR-bench](https://benchgap.net/calibration/babaisbench/current/to/ebrbench/current.md) | inverse Michaelis–Menten | 12 | 13.6 pp | 0 |
| [Terminal-Bench 2.1 (Vals AI) → AA Terminal-Bench 4.0](https://benchgap.net/calibration/vals-terminal-bench21/current/to/aa-terminal-bench4/current.md) | linear | 7 | 13.7 pp | 24 |
| [AA Cyber Index → BrowseComp](https://benchgap.net/calibration/aa-cyber-index/current/to/browsecomp/current.md) | Michaelis–Menten | 8 | 13.7 pp | 0 |
| [CountBench → SimpleVQA](https://benchgap.net/calibration/countbench/current/to/simplevqa/current.md) | Hill | 8 | 13.7 pp | 0 |
| [MazeBench → OSWorld 2.1 Offline](https://benchgap.net/calibration/mazebench/current/to/osworld-2-1-offline/current.md) | offset logistic | 11 | 13.8 pp | 6 |
| [NYT Connections → ExploitBench Internal Port](https://benchgap.net/calibration/nyt-connections/current/to/exploitbenchinternalport/current.md) | Hill | 6 | 13.8 pp | 9 |
| [Vals GPQA Diamond → ARC-AGI-2](https://benchgap.net/calibration/vals-gpqa-diamond/current/to/arc-agi-2/current.md) | linear | 31 | 13.8 pp | 22 |
| [Mystery Game Puzzles → NYT Connections](https://benchgap.net/calibration/mysterygamepuzzles/current/to/nyt-connections/current.md) | Hill | 54 | 13.8 pp | 5 |
| [CursorBench 3.2 → FrontierSWE v2](https://benchgap.net/calibration/cursor-bench32/current/to/frontierswev2/current.md) | Hill | 7 | 13.8 pp | 0 |
| [GPQA Diamond → SuperGPQA](https://benchgap.net/calibration/gpqa-diamond/current/to/supergpqa/current.md) | Hill | 7 | 13.9 pp | 12 |
| [Hermes Index → Agents' Last Exam](https://benchgap.net/calibration/hermesindex/current/to/agentslastexam/current.md) | linear | 7 | 13.9 pp | 0 |
| [HLE w/ tools → CyberGym](https://benchgap.net/calibration/hlewithtools/current/to/cybergym/current.md) | Michaelis–Menten | 7 | 14.0 pp | 0 |
| [DynaMath → VideoMMMU](https://benchgap.net/calibration/dynamath/current/to/videommmu/current.md) | offset logistic | 7 | 14.0 pp | 0 |
| [AA-SciCode → ScarfBench](https://benchgap.net/calibration/aa-sci-code/current/to/scarfbench/current.md) | inverse Michaelis–Menten | 10 | 14.0 pp | 18 |
| [Chess Puzzles → Furniture Assembly](https://benchgap.net/calibration/chesspuzzles/current/to/furnitureassembly/current.md) | offset logistic | 38 | 14.1 pp | 17 |
| [BioMysteryBench (human-difficult) → AA-Omniscience Accuracy](https://benchgap.net/calibration/biomysterybenchhumandifficult/current/to/omniscienceaccuracy/current.md) | Michaelis–Menten + offset | 6 | 14.1 pp | 0 |
| [ERQA → VideoMMMU](https://benchgap.net/calibration/erqa/current/to/videommmu/current.md) | Michaelis–Menten + offset | 8 | 14.1 pp | 0 |
| [WideResearch → GDPval-AA](https://benchgap.net/calibration/wideresearch/current/to/gdpvalaanormalized/current.md) | inverse Michaelis–Menten | 6 | 14.1 pp | 0 |
| [MaxxBench → BabaIsBench](https://benchgap.net/calibration/maxxbench/current/to/babaisbench/current.md) | Michaelis–Menten + offset | 9 | 14.1 pp | 4 |
| [ExploitBench Internal Port → MirrorCode](https://benchgap.net/calibration/exploitbenchinternalport/current/to/mirrorcode/current.md) | offset logistic | 7 | 14.2 pp | 2 |
| [BFCL v4 → MCP Atlas](https://benchgap.net/calibration/bfcl-v4/current/to/mcpatlas/current.md) | offset logistic | 7 | 14.2 pp | 5 |
| [MATH-500 → AIME 2025](https://benchgap.net/calibration/math-500/current/to/aime2025/current.md) | linear | 17 | 14.2 pp | 2 |
| [AA-SciCode → FrontierSWE v2](https://benchgap.net/calibration/aa-sci-code/current/to/frontierswev2/current.md) | Hill | 20 | 14.2 pp | 33 |
| [Terminal-Bench 4.0 → Terminal-Bench-Science 0.1](https://benchgap.net/calibration/terminal-bench/current/to/terminal-bench-science/current.md) | Hill | 11 | 14.2 pp | 10 |
| [AA Coding Index → FrontierSWE v2](https://benchgap.net/calibration/aa-coding-index/current/to/frontierswev2/current.md) | Hill | 13 | 14.3 pp | 7 |
| [CritPt → ARC-AGI-2](https://benchgap.net/calibration/critpt/current/to/arc-agi-2/current.md) | linear | 39 | 14.4 pp | 89 |
| [AA-HLE → MMLU-Pro](https://benchgap.net/calibration/aa-hle/current/to/mmlu-pro/current.md) | Hill | 50 | 14.4 pp | 1 |
| [AA-HLE → ARC-AGI-2](https://benchgap.net/calibration/aa-hle/current/to/arc-agi-2/current.md) | Hill | 38 | 14.4 pp | 12 |
| [HealthBench (length-adjusted) → ARC-AGI-3](https://benchgap.net/calibration/healthbenchlengthadjusted/current/to/arcagi3/current.md) | linear | 5 | 14.4 pp | 0 |
| [ResearchClawBench → JobBench](https://benchgap.net/calibration/researchclawbench/current/to/jobbench/current.md) | inverse Michaelis–Menten | 5 | 14.4 pp | 1 |
| [Gert Labs → MCP Atlas](https://benchgap.net/calibration/gertlabs/current/to/mcpatlas/current.md) | inverse Michaelis–Menten | 14 | 14.6 pp | 7 |
| [SWE-bench Verified → LiveCodeBench](https://benchgap.net/calibration/swe-bench-verified/current/to/livecodebench/current.md) | Michaelis–Menten | 5 | 14.6 pp | 107 |
| [NYT Connections → Mystery Game Puzzles](https://benchgap.net/calibration/nyt-connections/current/to/mysterygamepuzzles/current.md) | Hill | 54 | 14.6 pp | 13 |
| [Artificial Analysis Intelligence Index → ARC-AGI-2](https://benchgap.net/calibration/artificialanalysis/current/to/arc-agi-2/current.md) | Michaelis–Menten + offset | 40 | 14.6 pp | 7 |
| [HealthBench Professional (raw, Artificial Analysis) → AA-Omniscience Accuracy](https://benchgap.net/calibration/aa-health-bench-professional-raw/current/to/omniscienceaccuracy/current.md) | inverse Michaelis–Menten | 6 | 14.7 pp | 0 |
| [CWE-Bench-AA → OSWorld 2.0](https://benchgap.net/calibration/aa-cwe-bench/current/to/osworld2/current.md) | Hill | 7 | 14.7 pp | 8 |
| [DeepSearchQA → JobBench](https://benchgap.net/calibration/deepsearchqa/current/to/jobbench/current.md) | offset logistic | 9 | 14.8 pp | 1 |
| [DAYJOB Healthcare → EBR-bench](https://benchgap.net/calibration/dayjob-healthcare/current/to/ebrbench/current.md) | Hill | 11 | 14.8 pp | 0 |
| [ExploitBench → MirrorCode](https://benchgap.net/calibration/exploitbench/current/to/mirrorcode/current.md) | offset logistic | 9 | 14.9 pp | 7 |
| [Terminal-Bench 2.1 (Vals AI) → Terminal-Bench 4.1](https://benchgap.net/calibration/vals-terminal-bench21/current/to/terminal-bench41/current.md) | inverse Michaelis–Menten | 32 | 14.9 pp | 18 |
| [SEC-Bench Pro → MirrorCode](https://benchgap.net/calibration/secbenchpro/current/to/mirrorcode/current.md) | inverse Michaelis–Menten | 10 | 15.0 pp | 1 |
