General
FACTS Multimodal leaderboard
As of 2026-11-18, the highest measured score on FACTS Multimodal is 60.2% by Gemini 4 Carbon. 102 more models have estimated scores, calibrated from the benchmarks they were measured on.
Measured scores: benchlm.ai.
| # | Model | Score | Source |
|---|---|---|---|
| 1 | Gemini 4 Carbon | 60.2% | measured |
| 2 | Gemini 3.8 Flash | 57.6% | measured |
| 3 | Claude Opus 5.5 | 56.4% | measured |
| 4 | GPT-6 Astra | 55.2% | measured |
| 5 | Claude Fable 5.2 | 54.2% | estimated ± 4.0 pp, high confidence |
| 6 | Gemini 4 Argon | 54.2% | measured |
| 7 | Gemini 3.6 Flash | 53.7% | measured |
| 8 | Gemini 3 Pro | 52.7% | estimated ± 4.0 pp, high confidence |
| 9 | Claude Opus 5 | 52.6% | estimated ± 4.0 pp, high confidence |
| 10 | GPT-6.1 Sol | 52.6% | measured |
| 11 | Gemini 3.7 Flash | 52.5% | estimated ± 4.0 pp, high confidence |
| 12 | GPT-6.2 Astra | 52.4% | estimated ± 2.7 pp, medium confidence |
| 13 | GPT-6.2 Sol | 52.4% | estimated ± 2.7 pp, medium confidence |
| 14 | GPT-6 Cyber | 52.4% | estimated ± 2.7 pp, medium confidence |
| 15 | Muse Wonder | 52.2% | estimated ± 4.0 pp, high confidence |
| 16 | Claude Fable 5 | 51.8% | estimated ± 4.0 pp, high confidence |
| 17 | Gemini 3.5 Flash | 51.7% | estimated ± 4.0 pp, high confidence |
| 18 | GPT-5.5 Pro | 51.6% | estimated ± 5.6 pp, medium confidence |
| 19 | Qwen3.8 Max | 51.5% | measured |
| 20 | Claude Opus 4.6 (Adaptive) | 51.5% | estimated ± 4.0 pp, high confidence |
| 21 | Claude Mythos 5.2 | 51.2% | estimated ± 2.7 pp, medium confidence |
| 22 | GPT-5.4 | 51.1% | estimated ± 4.0 pp, high confidence |
| 23 | Claude Opus 4.8 | 50.9% | estimated ± 4.0 pp, high confidence |
| 24 | GPT-5.4 Pro | 50.5% | estimated ± 5.6 pp, medium confidence |
| 25 | Claude Sonnet 5.5 | 50.4% | measured |
| 26 | Gemini 3.1 Pro | 50.4% | measured |
| 27 | Claude Fable 5.1 | 50.4% | estimated ± 4.0 pp, high confidence |
| 28 | Grok 4.20 Multi-agent | 50.2% | estimated ± 4.0 pp, high confidence |
| 29 | Gemini 3 Flash | 49.7% | measured |
| 30 | Grok 4.1 Fast (Reasoning) | 49.0% | estimated ± 4.0 pp, high confidence |
| 31 | Kimi K2.6 | 48.9% | estimated ± 4.0 pp, high confidence |
| 32 | GPT-6.1 Terra | 48.8% | measured |
| 33 | Grok 4.20 | 48.0% | estimated ± 4.0 pp, high confidence |
| 34 | Muse Spark 1.3 | 47.8% | estimated ± 4.0 pp, high confidence |
| 35 | Qwen3.7 Max | 47.8% | estimated ± 4.0 pp, high confidence |
| 36 | Muse Spark 1.1 | 47.7% | estimated ± 4.0 pp, high confidence |
| 37 | Claude Mythos 5 | 47.2% | estimated ± 2.7 pp, medium confidence |
| 38 | GPT-5.2 | 47.0% | estimated ± 4.0 pp, high confidence |
| 39 | Hy4 | 47.0% | estimated ± 4.0 pp, high confidence |
| 40 | GPT-5.6 Sol | 46.4% | measured |
| 41 | Kimi K3.1 Pro | 46.4% | estimated ± 2.7 pp, medium confidence |
| 42 | GPT-5.4 nano | 45.4% | estimated ± 5.6 pp, medium confidence |
| 43 | Grok 4.5 | 45.1% | estimated ± 4.0 pp, high confidence |
| 44 | Claude Mythos Preview | 45.0% | estimated ± 2.7 pp, medium confidence |
| 45 | Muse Spark 1.2 | 44.7% | estimated ± 4.0 pp, high confidence |
| 46 | Muse Spark Open | 44.3% | estimated ± 4.0 pp, high confidence |
| 47 | GPT-5.1 | 44.3% | estimated ± 5.6 pp, medium confidence |
| 48 | GLM-5.1 | 43.9% | estimated ± 4.0 pp, high confidence |
| 49 | GPT-6 Sol | 43.9% | measured |
| 50 | Claude Opus 4.7 | 43.9% | estimated ± 5.6 pp, medium confidence |
| 51 | Step 5 Preview | 43.3% | estimated ± 4.0 pp, high confidence |
| 52 | DeepSeek V4.1 Flash 1017 | 42.6% | estimated ± 3.0 pp, high confidence |
| 53 | Claude Sonnet 5 | 42.5% | estimated ± 4.0 pp, high confidence |
| 54 | Qwen3.7 Plus | 42.4% | estimated ± 4.0 pp, high confidence |
| 55 | Gemma 4 31B | 42.3% | measured |
| 56 | GLM-5.2 | 42.1% | estimated ± 4.0 pp, high confidence |
| 57 | GPT-5 | 42.1% | estimated ± 5.6 pp, medium confidence |
| 58 | o3 | 42.1% | estimated ± 5.6 pp, medium confidence |
| 59 | Qwen 3.6 Max (preview) | 42.0% | estimated ± 4.0 pp, high confidence |
| 60 | DeepSeek V4 Pro | 41.7% | estimated ± 4.7 pp, high confidence |
| 61 | Ling 3.1 Flash | 41.6% | estimated ± 4.7 pp, high confidence |
| 62 | GLM-5.3 | 41.5% | estimated ± 2.7 pp, medium confidence |
| 63 | Qwen3.5 397B | 41.5% | estimated ± 5.8 pp, low confidence |
| 64 | Grok 4.7 | 40.9% | measured |
| 65 | DeepSeek V4 Flash 0731 | 40.6% | estimated ± 5.6 pp, medium confidence |
| 66 | MiMo-V2.6-Pro | 40.0% | estimated ± 2.7 pp, medium confidence |
| 67 | GPT-5.6 Terra | 39.9% | measured |
| 68 | Qwen3.6-27B | 39.8% | estimated ± 4.7 pp, high confidence |
| 69 | Beam 1.0 501B-A23B | 39.7% | estimated ± 4.7 pp, high confidence |
| 70 | GPT-5.5 | 39.5% | estimated ± 3.0 pp, high confidence |
| 71 | Gemma 4 26B A4B | 39.5% | measured |
| 72 | Inkling-Small | 39.5% | estimated ± 4.7 pp, high confidence |
| 73 | Grok 4.6 | 39.3% | measured |
| 74 | Nemotron 3 Ultra | 39.3% | estimated ± 4.7 pp, medium confidence |
| 75 | MAI-Thinking-1 | 39.3% | estimated ± 4.7 pp, medium confidence |
| 76 | Hy4 preview | 38.9% | estimated ± 4.0 pp, high confidence |
| 77 | Claude Haiku 5.5 | 38.9% | measured |
| 78 | GPT-6 Luna | 38.7% | measured |
| 79 | Qwen3 Max | 38.6% | estimated ± 5.8 pp, low confidence |
| 80 | Seed 2.1 Turbo | 38.4% | estimated ± 5.6 pp, medium confidence |
| 81 | GPT-OSS 120B | 37.5% | estimated ± 5.8 pp, low confidence |
| 82 | GPT-5.6 Luna | 37.4% | measured |
| 83 | MiniMax M3 | 37.3% | estimated ± 4.0 pp, high confidence |
| 84 | SWE-2 Flash | 36.9% | estimated ± 2.7 pp, medium confidence |
| 85 | GLM-4.7 | 36.6% | estimated ± 5.6 pp, medium confidence |
| 86 | GLM-5.3-Flash | 36.6% | estimated ± 5.6 pp, medium confidence |
| 87 | Inkling | 36.6% | estimated ± 5.6 pp, medium confidence |
| 88 | Seed 2.1 Pro | 36.6% | estimated ± 5.6 pp, medium confidence |
| 89 | Kimi K3 | 36.2% | estimated ± 2.7 pp, low confidence |
| 90 | Gemini 3.5 Flash-Lite | 36.1% | measured |
| 91 | GPT-5.4 mini | 35.5% | estimated ± 4.0 pp, high confidence |
| 92 | Qwen3.6 Plus | 34.7% | estimated ± 4.0 pp, medium confidence |
| 93 | MiMo-V2.6-Flash | 34.6% | estimated ± 2.7 pp, low confidence |
| 94 | o3-mini | 34.4% | estimated ± 5.6 pp, low confidence |
| 95 | Kimi K2.7 Code | 34.0% | estimated ± 5.6 pp, low confidence |
| 96 | Qwen3.5 Flash | 34.0% | estimated ± 5.6 pp, low confidence |
| 97 | o4-mini | 33.6% | estimated ± 5.6 pp, low confidence |
| 98 | GPT-4.1 | 32.9% | estimated ± 5.6 pp, low confidence |
| 99 | o1 | 32.5% | estimated ± 5.6 pp, low confidence |
| 100 | GPT-4o | 32.2% | estimated ± 5.6 pp, low confidence |
| 101 | Qwen3.5 Plus | 32.2% | estimated ± 5.6 pp, low confidence |
| 102 | Grok 4.3 | 32.1% | estimated ± 4.0 pp, medium confidence |
| 103 | Claude Haiku 4.5 | 31.8% | estimated ± 5.6 pp, low confidence |
| 104 | Qwen3.8-27B | 31.7% | estimated ± 4.0 pp, medium confidence |
| 105 | Claude Opus 4.5 Thinking | 30.6% | estimated ± 4.0 pp, medium confidence |
| 106 | Claude 3 Opus | 30.3% | estimated ± 5.6 pp, low confidence |
| 107 | Qwen3.5-122B-A10B | 30.2% | estimated ± 4.0 pp, medium confidence |
| 108 | GPT-4o mini | 30.0% | estimated ± 5.6 pp, low confidence |
| 109 | GPT-4.1 mini | 29.2% | estimated ± 5.6 pp, low confidence |
| 110 | Claude Opus 4.5 | 29.0% | estimated ± 4.0 pp, medium confidence |
| 111 | GPT-4.1 nano | 28.9% | estimated ± 5.6 pp, low confidence |
| 112 | Qwen3.5-27B | 28.2% | estimated ± 4.0 pp, medium confidence |
| 113 | Claude Sonnet 4.6 | 26.6% | estimated ± 4.0 pp, medium confidence |
| 114 | Qwen3.7 Flash | 26.1% | estimated ± 4.0 pp, medium confidence |
| 115 | Claude Opus 4.6 | 25.8% | estimated ± 4.0 pp, medium confidence |
| 116 | Qwen3.6-35B-A3B | 24.9% | estimated ± 4.0 pp, medium confidence |
| 117 | Hy3 | 24.7% | estimated ± 4.0 pp, medium confidence |
| 118 | Step 3.7 Flash | 23.9% | estimated ± 4.0 pp, medium confidence |
| 119 | Claude Sonnet 4.5 Thinking | 22.7% | estimated ± 4.0 pp, medium confidence |
| 120 | Claude Sonnet 4.5 | 21.9% | estimated ± 4.0 pp, medium confidence |
| 121 | MiMo-V2.5-Pro | 21.1% | estimated ± 4.0 pp, medium confidence |
| 122 | MiniMax M3.1 Flash | 17.9% | estimated ± 4.0 pp, medium confidence |
| 123 | Mistral Large 4 | 17.4% | estimated ± 4.0 pp, medium confidence |
| 124 | MiMo-V2-Pro | 16.6% | estimated ± 4.0 pp, medium confidence |
| 125 | MiniMax M2.7 | 16.0% | estimated ± 4.0 pp, medium confidence |
All leaderboards
Agentic · terminal
- AA Terminal-Bench 2.1
- AA Terminal-Bench 4.0
- Terminal-Bench 4.0
- Terminal-Bench 2.0
- Terminal-Bench 2.1
- Terminal-Bench 3.0
- Terminal-Bench Hard
- Terminal-Bench-Science 0.1
- Terminal-Bench 4.1
- Terminal-Bench 2.1 (Vals AI)
Agentic · tools
- AA Agentic Index
- AA-AnalystAgent
- AA AutomationBench
- CWE-Bench-AA
- AA Cyber Index
- DeepsecBench-AA
- AA EnterpriseOps-Gym
- GDP.pdf
- AA Harvey LAB v1.0
- Harvey LAB-AA v1.1
- AA ITBench
- AA Tau3 Banking
- Agents' Last Exam
- APEX-Agents
- APEX-Agents-AA
- ApprenticeBench
- AutomationBench
- BFCL v4
- BrowseComp
- Claw-Eval
- CWE-bench v1
- CWE-bench
- CyberGym
- DeepPlanning
- DeepSearchQA
- DRACO
- ExploitGym
- Finance Agent v2
- GDPval-AA
- Gert Labs
- Hermes Bench
- Hermes Index
- HLE w/ tools
- JobBench
- LifeVal
- MazeBench
- MCP Atlas
- MCP-Tasks
- MKA
- OSWorld 2.1
- OSWorld 2.1 Offline
- OSWorld-Verified
- OSWorld 2.0
- PinchBench
- QwenClawBench
- ResearchClawBench
- SkillsBench
- τ²-bench results
- τ³-bench results
- Toolathlon
- Toolathlon-Verified
- VITA-Bench
- WideResearch
Coding
- AA Coding Index
- AA LiveCodeBench
- AA-SciCode
- Bug Hunt Bench
- CursorBench 3.1
- CursorBench 3.2
- CursorBench 4.0
- DeepSWE
- FrontierCode 1.1 Main
- FrontierCode 2.0
- FrontierCode 1.1 Extended
- FrontierSWE v2
- LiveCodeBench
- LiveCodeBench Pro
- LiveCodeBench v6
- NL2Repo
- OpenHarmony Bench v1.2
- OpenHarmony Bench v1.0
- PostTrainBench
- PostTrainBench v1
- PostTrainBench v1.1
- ProgramBench
- React Native Evals
- ScarfBench
- SciCode
- SWE-Atlas Codebase QnA
- SWE-bench Pro
- SWE-bench Verified
- SWE-Rebench
- SWE-Marathon
- Vals LiveCodeBench
- Vals SWE-bench
- Vibe Code Bench
- VulcanBench Frontier v4
- VulcanBench v3
Math
- AIME 2024
- AIME 2025
- AIME26
- FrontierMath v2 (Tier 4)
- FrontierMath v2 (Tiers 1-3)
- HMMT Feb 2025
- HMMT Feb 2026
- HMMT Nov 2025
- IMOAnswerBench
- MATH-500
- MMAnswerBench
Knowledge & reasoning
- AA-GPQA Diamond
- HealthBench Professional (length-adjusted, Artificial Analysis)
- HealthBench Professional (raw, Artificial Analysis)
- AA-HLE
- AA MMLU-Pro
- ARC-AGI-2
- ARC-AGI-3 Provider Adapter
- ARC-AGI-1
- ARC-AGI-3
- Artificial Analysis Intelligence Index
- BBH
- BioBasicsBench
- BioMysteryBench (human-difficult)
- BioMysteryBench (human-solvable)
- C-Eval
- CritPt
- GPQA
- GPQA Diamond
- HealthBench (raw)
- HealthBench Hard
- HealthBench (length-adjusted)
- HealthBench Professional
- HealthBench Professional (raw)
- HLE
- HLE w/o tools
- HLE-Verified
- JevBench 1.4
- LABBench2
- MedXpertQA (Text)
- MMLU
- MMLU-Pro
- MMLU-Pro (Arcee)
- MMLU-Redux
- MMMLU
- AA-Omniscience Accuracy
- SimpleQA
- STEMBench
- SuperGPQA
- Vals GPQA Diamond
- Vals MMLU-Pro
Instruction following
Multilingual
Vision & documents
- AA-MMMU-Pro
- AI2D_TEST
- BabyVision
- BabyVision w/ Python
- BenchCAD Vision2Code (no tools)
- BenchCAD Vision2Code (tools)
- CC-OCR
- Chartography (no tools)
- Chartography (tools)
- CharXiv
- CharXiv w/o tools
- CountBench
- DynaMath
- ERQA
- LVBench
- MathVision
- MathVision w/ Python
- MedXpertQA (MM)
- MLVU (M-Avg)
- MMMU
- MMMU-Pro
- MMMU-Pro w/ Python
- MMVU
- OCRBench V2
- OfficeQA
- OfficeQA Pro
- OmniDocBench 1.5
- ParseBench-1
- ParseBench-2
- PerceptionBench
- RealWorldQA
- RefCOCO (avg)
- ScreenSpot Pro
- SimpleVQA
- SWE-bench Multimodal
- Video-MME (w/o subtitle)
- Video-MME (with subtitle)
- VideoMMMU
- V*
- We-Math
- ZeroBench
Long context
General
- HLE-MultiChoice (Decision Index)
- Artificial Analysis Decision Index v1.0
- Artificial Analysis Decision Index
- When2Call (Decision Index)
- WildGuardTest (Decision Index)
- BabaIsBench
- BinaryAudit
- BlenderWorld (text-only)
- Chess Puzzles
- CompileBench
- Cookie-Bench
- DAYJOB
- DAYJOB Healthcare
- Decision Index Full Score
- EBR-bench
- Epoch Automation Reports
- ExploitBench
- ExploitBench Internal Port
- FACTS Multimodal
- FailureSensorIQ
- FrontierPhysics
- Furniture Assembly
- Decision Index Arts
- Decision Index Knowledge
- Decision Index Language
- Decision Index Retrieval
- Decision Index Tools
- MaxxBench
- MirrorCode
- Mystery Game Puzzles
- NYT Connections
- OTelBench
- SEC-Bench Pro
- SkillsBench official leaderboard (BenchFlow)
- SpireClimb
- StealthMark