Agentic · tools
AutomationBench leaderboard
As of 2026-11-21, the highest measured score on AutomationBench is 62.2% by MiniMax M3.1 Flash. 69 more models have estimated scores, calibrated from the benchmarks they were measured on.
Measured scores: benchlm.ai.
| # | Model | Score | Source |
|---|---|---|---|
| 1 | MiniMax M3.1 Flash | 62.2% | measured |
| 2 | ERNIE 5.5 Flash Preview | 61.9% | measured |
| 3 | DeepSeek V4.1 Flash 1017 | 60.5% | measured |
| 4 | Hy4 | 56.2% | measured |
| 5 | Gemini 4 Carbon | 53.9% | measured |
| 6 | Atria Dawn Preview | 53.8% | measured |
| 7 | Gemini 4 Neon | 53.4% | estimated ± 6.1 pp, low confidence |
| 8 | MiMo-V2.6-Pro | 53.1% | measured |
| 9 | Ling 3.1 Flash | 52.5% | measured |
| 10 | MiMo-V2.6-Flash | 52.3% | measured |
| 11 | GPT-6.2 Astra | 52.3% | estimated ± 6.1 pp, low confidence |
| 12 | Hy4 flash | 52.2% | measured |
| 13 | LongCat-2.5 | 51.8% | measured |
| 14 | Claude Fable 5.2 | 51.4% | measured |
| 15 | Gemini 4 Argon | 51.3% | measured |
| 16 | Gemini 4 Hydrogen | 50.6% | estimated ± 6.1 pp, low confidence |
| 17 | Ultramarine 1 Flash | 50.2% | measured |
| 18 | Mercury 2.5 | 49.9% | estimated ± 9.1 pp, low confidence |
| 19 | Mistral Medium 3.5 128B | 49.7% | estimated ± 9.1 pp, low confidence |
| 20 | Muse Spark 1.3 | 49.4% | measured |
| 21 | Fugu Cyber | 49.2% | estimated ± 10.4 pp, low confidence |
| 22 | GLM-5.3-Flash | 48.8% | measured |
| 23 | MiMo-V2.5-Pro | 48.8% | estimated ± 9.1 pp, low confidence |
| 24 | Nemotron 3 Ultra | 48.6% | estimated ± 9.1 pp, low confidence |
| 25 | Qwen3.6 Plus | 48.6% | estimated ± 9.1 pp, low confidence |
| 26 | GLM-5.1 | 48.6% | estimated ± 9.1 pp, low confidence |
| 27 | Claude Opus 4.5 | 48.6% | estimated ± 9.1 pp, low confidence |
| 28 | Grok 4.6 | 48.5% | estimated ± 7.1 pp, low confidence |
| 29 | Qwen3.5 397B | 48.5% | estimated ± 9.1 pp, low confidence |
| 30 | MiniMax M2.7 | 48.4% | estimated ± 9.1 pp, low confidence |
| 31 | Qwen3.6-35B-A3B | 48.4% | estimated ± 9.1 pp, low confidence |
| 32 | Pokee-Isaac 28B | 48.3% | estimated ± 9.1 pp, low confidence |
| 33 | Gemini 3.8 Flash Cyber | 48.3% | estimated ± 10.4 pp, low confidence |
| 34 | Kimi K2.5 | 48.3% | estimated ± 9.1 pp, low confidence |
| 35 | GLM-5 | 48.3% | estimated ± 9.1 pp, low confidence |
| 36 | GLM-5.3 | 48.2% | measured |
| 37 | Granite 4.2 30B | 48.0% | estimated ± 9.1 pp, low confidence |
| 38 | Granite 4.2 8B | 47.6% | estimated ± 9.1 pp, low confidence |
| 39 | MiniMax M3 | 47.6% | estimated ± 9.5 pp, low confidence |
| 40 | Kimi K3.1 Pro | 47.4% | estimated ± 6.1 pp, low confidence |
| 41 | Qwen4 Max | 47.3% | estimated ± 6.1 pp, low confidence |
| 42 | Claude Sonnet 5.5 | 47.2% | estimated ± 6.1 pp, low confidence |
| 43 | GPT-6.2 Sol | 47.1% | estimated ± 6.1 pp, low confidence |
| 44 | Muse Ember | 47.1% | estimated ± 6.1 pp, low confidence |
| 45 | Grok 4.7 | 47.1% | estimated ± 6.1 pp, low confidence |
| 46 | GPT-6.1 Terra | 47.1% | estimated ± 6.1 pp, low confidence |
| 47 | Mistral Large 4 | 47.1% | estimated ± 6.1 pp, low confidence |
| 48 | GPT-6 Luna | 47.1% | estimated ± 6.1 pp, low confidence |
| 49 | Beam 1.0 501B-A23B | 47.1% | estimated ± 6.1 pp, low confidence |
| 50 | Qwen3.8-27B | 47.1% | estimated ± 6.1 pp, low confidence |
| 51 | Inkling 2 Small | 47.1% | estimated ± 6.1 pp, low confidence |
| 52 | Muse Wonder Full | 47.1% | estimated ± 6.1 pp, low confidence |
| 53 | Granite 4.2 3B | 46.2% | estimated ± 9.1 pp, low confidence |
| 54 | Claude Mythos 5 | 45.1% | estimated ± 10.4 pp, low confidence |
| 55 | Beam 1.0 501B-A23B FP8 | 44.7% | estimated ± 9.1 pp, low confidence |
| 56 | Beam 1.0 501B-A23B NVFP4 | 44.4% | estimated ± 9.1 pp, low confidence |
| 57 | Spark-X2.6-4B | 44.2% | estimated ± 9.1 pp, low confidence |
| 58 | Gemini 3.5 Flash Cyber | 44.2% | estimated ± 10.4 pp, low confidence |
| 59 | Claude Mythos Preview | 44.1% | estimated ± 10.4 pp, low confidence |
| 60 | Step 5 Preview | 44.0% | measured |
| 61 | Mercury 3 | 43.1% | estimated ± 9.1 pp, low confidence |
| 62 | GPT-6 Astra | 41.4% | measured |
| 63 | LFM3-1.7B | 41.3% | measured |
| 64 | Claude Opus 5.5 | 40.0% | measured |
| 65 | Gemini 3.8 Flash | 39.2% | estimated ± 9.5 pp, low confidence |
| 66 | Nemotron 3.5 Super VL 120B A12B | 38.5% | estimated ± 9.1 pp, low confidence |
| 67 | Nemotron 3.5 Super VL 120B A12B NVFP4 | 37.9% | estimated ± 9.1 pp, low confidence |
| 68 | Nemotron 3.5 Bolt 74B-A9B | 37.4% | estimated ± 9.1 pp, low confidence |
| 69 | Nemotron 3.5 Bolt 74B-A9B BF16 | 37.4% | estimated ± 9.1 pp, low confidence |
| 70 | Apodex 1.1 | 37.1% | estimated ± 7.4 pp, low confidence |
| 71 | GPT-5.6 Luna | 36.5% | estimated ± 10.4 pp, low confidence |
| 72 | Ripple-65B Preview | 36.4% | measured |
| 73 | GPT-6.1 Sol | 36.1% | measured |
| 74 | Qwen3.8-Flash-Next | 36.1% | estimated ± 7.1 pp, low confidence |
| 75 | GPT-5.6 Sol | 34.0% | estimated ± 7.1 pp, low confidence |
| 76 | GPT-6 Sol | 33.2% | measured |
| 77 | Hy4 preview | 32.1% | measured |
| 78 | DeepSeek V4 Pro | 31.8% | measured |
| 79 | Claude Fable 5.1 | 31.4% | measured |
| 80 | dots3-note Preview | 31.0% | estimated ± 7.4 pp, low confidence |
| 81 | Kimi K3 | 30.8% | measured |
| 82 | Gemini 3.7 Flash | 30.4% | measured |
| 83 | Grok 4.5 | 30.2% | estimated ± 7.1 pp, low confidence |
| 84 | Nemotron 3.5 Lightning 30B A3B NVFP4 | 30.1% | estimated ± 9.1 pp, low confidence |
| 85 | Claude Opus 4.7 | 29.2% | estimated ± 10.4 pp, low confidence |
| 86 | Apodex 1.1 Mini | 28.4% | estimated ± 7.4 pp, low confidence |
| 87 | LFM2.5-24B-A2B | 27.7% | measured |
| 88 | GPT-5.6 Terra | 27.4% | estimated ± 7.1 pp, low confidence |
| 89 | Qwen3.8 Max | 27.3% | measured |
| 90 | GPT-5.4 | 26.5% | estimated ± 7.1 pp, low confidence |
| 91 | Claude Opus 5 | 26.0% | measured |
| 92 | GPT-5.5 | 25.7% | estimated ± 7.1 pp, low confidence |
| 93 | Ling 3.0 Flash Fin | 25.2% | estimated ± 7.1 pp, low confidence |
| 94 | DeepSeek V4 Flash 0731 | 25.1% | measured |
| 95 | Claude Fable 5 | 24.5% | estimated ± 7.1 pp, low confidence |
| 96 | Claude Sonnet 5 | 23.5% | estimated ± 7.1 pp, low confidence |
| 97 | LFM2.5-2.6B | 23.2% | estimated ± 9.1 pp, low confidence |
| 98 | Solar Pro 4 | 20.2% | estimated ± 7.4 pp, low confidence |
| 99 | Claude Opus 4.6 | 19.8% | estimated ± 10.4 pp, low confidence |
| 100 | Solar Open 2 | 18.2% | estimated ± 7.4 pp, low confidence |
| 101 | Claude Sonnet 4.6 | 18.0% | estimated ± 10.4 pp, low confidence |
| 102 | Muse Spark 1.1 | 11.0% | estimated ± 10.4 pp, low 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
- cursorBench41
- 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
- ZeroBench w/ Python
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
- 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