Coding
FrontierSWE v2 leaderboard
As of 2026-11-21, the highest measured score on FrontierSWE v2 is 69.9% by Claude Fable 5.2. 154 more models have estimated scores, calibrated from the benchmarks they were measured on.
Measured scores: benchlm.ai.
| # | Model | Score | Source |
|---|---|---|---|
| 1 | Claude Fable 5.2 | 69.9% | measured |
| 2 | GPT-6 Astra | 65.5% | measured |
| 3 | Gemini 4 Neon | 62.4% | measured |
| 4 | Claude Opus 5.5 | 62.3% | measured |
| 5 | Claude Sonnet 5.5 | 61.9% | measured |
| 6 | Gemini 4 Carbon | 61.2% | measured |
| 7 | GPT-6.2 Astra | 60.6% | estimated ± 8.0 pp, medium confidence |
| 8 | GPT-6.2 Sol | 60.2% | estimated ± 8.0 pp, medium confidence |
| 9 | Claude Fable 5.1 | 56.3% | measured |
| 10 | GPT-6.1 Sol | 56.0% | estimated ± 8.0 pp, medium confidence |
| 11 | Gemini 4 Argon | 55.0% | measured |
| 12 | Gemini 4 Hydrogen | 53.1% | measured |
| 13 | Claude Opus 5 | 52.0% | measured |
| 14 | GPT-6 Sol | 51.1% | estimated ± 8.0 pp, medium confidence |
| 15 | MiniMax M3.1 Flash | 50.8% | estimated ± 10.4 pp, low confidence |
| 16 | Muse Wonder Full | 50.7% | estimated ± 11.1 pp, low confidence |
| 17 | Claude Mythos 5 | 50.7% | estimated ± 8.3 pp, medium confidence |
| 18 | Muse Wonder | 47.6% | estimated ± 5.8 pp, medium confidence |
| 19 | Claude Fable 5 | 47.0% | measured |
| 20 | Hy4 | 46.7% | estimated ± 8.0 pp, medium confidence |
| 21 | Claude Haiku 5.5 | 43.8% | measured |
| 22 | Qwen4 Max | 42.3% | estimated ± 11.2 pp, low confidence |
| 23 | DeepSeek V4.1 Flash 1017 | 42.0% | estimated ± 5.8 pp, medium confidence |
| 24 | Sakana Fugu-Ultra | 39.3% | estimated ± 8.3 pp, medium confidence |
| 25 | Step 5 Preview | 39.2% | estimated ± 14.2 pp, low confidence |
| 26 | SWE-2 | 38.3% | estimated ± 8.0 pp, medium confidence |
| 27 | SWE-2 Flash | 38.2% | estimated ± 8.0 pp, medium confidence |
| 28 | GPT-5.5 | 37.6% | estimated ± 8.0 pp, medium confidence |
| 29 | GPT-6 Luna | 36.5% | estimated ± 8.0 pp, medium confidence |
| 30 | Kimi K3.1 Pro | 32.7% | estimated ± 8.0 pp, medium confidence |
| 31 | Mistral Large 4 | 32.4% | estimated ± 8.0 pp, medium confidence |
| 32 | GPT-5.6 Sol | 32.2% | measured |
| 33 | MiMo-V2.6-Pro | 31.0% | estimated ± 11.1 pp, low confidence |
| 34 | Ling 3.1 Flash | 30.7% | estimated ± 8.3 pp, medium confidence |
| 35 | GLM-5.3 | 30.2% | measured |
| 36 | Grok 4.7 | 29.5% | measured |
| 37 | GPT-6.1 Terra | 29.5% | estimated ± 14.2 pp, low confidence |
| 38 | GPT-5.6 Terra | 28.6% | estimated ± 5.8 pp, medium confidence |
| 39 | SWE-1.7 | 26.4% | estimated ± 8.0 pp, low confidence |
| 40 | Claude Opus 4.8 | 26.1% | estimated ± 8.0 pp, low confidence |
| 41 | Beam 1.0 501B-A23B | 26.1% | estimated ± 8.3 pp, medium confidence |
| 42 | Kimi K3 | 25.9% | measured |
| 43 | Inkling 2 Mini | 25.8% | estimated ± 8.0 pp, low confidence |
| 44 | Muse Spark 1.3 | 25.8% | measured |
| 45 | Beam 1.0 501B-A23B FP8 | 25.8% | estimated ± 8.3 pp, medium confidence |
| 46 | Hy4 preview | 25.4% | estimated ± 8.3 pp, medium confidence |
| 47 | Grok 4.6 | 25.3% | measured |
| 48 | Beam 1.0 501B-A23B NVFP4 | 25.3% | estimated ± 8.3 pp, medium confidence |
| 49 | Inkling 2 Small | 25.1% | measured |
| 50 | Ornith-1.5-397B | 24.4% | estimated ± 8.3 pp, medium confidence |
| 51 | Grok 4.5 | 23.7% | estimated ± 8.3 pp, medium confidence |
| 52 | Claude Opus 4.7 | 23.0% | estimated ± 8.3 pp, medium confidence |
| 53 | Qwen3.8 Max Preview | 22.6% | estimated ± 14.2 pp, low confidence |
| 54 | Qwen3.8-Omni-Flash | 21.3% | estimated ± 8.3 pp, medium confidence |
| 55 | MiMo-V2.6-Flash | 21.0% | estimated ± 14.2 pp, low confidence |
| 56 | Gemini 3.7 Flash | 20.3% | measured |
| 57 | Qwen3.8-Flash-Next | 19.9% | estimated ± 8.3 pp, medium confidence |
| 58 | Gemini 3.8 Flash | 19.6% | measured |
| 59 | Ornith-1.0-397B | 19.4% | estimated ± 8.3 pp, medium confidence |
| 60 | GLM-5.2 | 19.2% | estimated ± 8.3 pp, medium confidence |
| 61 | Qwen3.8-27B | 18.5% | estimated ± 8.3 pp, medium confidence |
| 62 | Muse Spark 1.1 | 18.2% | estimated ± 8.3 pp, medium confidence |
| 63 | GLM-5.3-Flash | 18.1% | measured |
| 64 | Ripple-65B Preview | 18.0% | estimated ± 8.3 pp, medium confidence |
| 65 | Sage-A1.5 | 17.7% | estimated ± 8.3 pp, medium confidence |
| 66 | GPT-5.6 Luna | 17.3% | estimated ± 5.8 pp, low confidence |
| 67 | dots3-note Preview | 17.3% | estimated ± 8.3 pp, medium confidence |
| 68 | Qwen3.7 Max | 16.6% | estimated ± 8.3 pp, medium confidence |
| 69 | Hy3 | 15.9% | estimated ± 14.2 pp, low confidence |
| 70 | Hy3 Preview | 15.9% | estimated ± 14.2 pp, low confidence |
| 71 | Qwen3.8 Max | 15.8% | measured |
| 72 | Quasar 438B | 15.1% | estimated ± 14.2 pp, low confidence |
| 73 | Solar Pro 4 | 14.9% | estimated ± 14.2 pp, low confidence |
| 74 | Atria Dawn Preview | 14.9% | estimated ± 8.3 pp, medium confidence |
| 75 | Ornith-1.5-35B-A3B | 14.9% | estimated ± 8.3 pp, medium confidence |
| 76 | Laguna S 2.1 | 14.5% | estimated ± 8.3 pp, medium confidence |
| 77 | MiniMax M3 | 13.9% | estimated ± 8.3 pp, medium confidence |
| 78 | Sakana Fugu | 13.9% | estimated ± 8.3 pp, medium confidence |
| 79 | Claude Sonnet 5 | 13.6% | estimated ± 5.8 pp, low confidence |
| 80 | Kimi K2.6 | 13.2% | estimated ± 8.3 pp, medium confidence |
| 81 | GLM-5.1 | 12.8% | estimated ± 8.3 pp, medium confidence |
| 82 | Gemini 2.5 Pro | 12.3% | estimated ± 14.2 pp, low confidence |
| 83 | Muse Spark Open | 12.0% | estimated ± 8.0 pp, low confidence |
| 84 | Muse Spark 1.2 | 12.0% | measured |
| 85 | GPT-5.4 | 11.6% | estimated ± 8.3 pp, medium confidence |
| 86 | Qwen3.7 Plus | 11.4% | estimated ± 8.3 pp, medium confidence |
| 87 | Apodex 1.1 | 11.2% | estimated ± 14.2 pp, low confidence |
| 88 | Apodex 1.1 Mini | 11.2% | estimated ± 14.2 pp, low confidence |
| 89 | Gemma 4 31B | 11.2% | estimated ± 14.2 pp, low confidence |
| 90 | Qwen 3.6 Max (preview) | 10.9% | estimated ± 8.3 pp, medium confidence |
| 91 | MiMo-V2.5-Pro | 10.7% | estimated ± 8.3 pp, medium confidence |
| 92 | Claude Opus 4.5 | 10.6% | estimated ± 8.3 pp, medium confidence |
| 93 | Gemini 3.6 Flash | 10.1% | estimated ± 9.5 pp, medium confidence |
| 94 | GPT-5.3 Codex | 10.1% | estimated ± 8.3 pp, medium confidence |
| 95 | Ling 3.0 Flash | 9.7% | estimated ± 8.3 pp, medium confidence |
| 96 | Qwen3.6 Plus | 9.7% | estimated ± 8.3 pp, medium confidence |
| 97 | Gemini 3.1 Pro | 9.7% | estimated ± 9.5 pp, medium confidence |
| 98 | Ling 3.0 Flash VL | 9.6% | estimated ± 14.2 pp, low confidence |
| 99 | Claude Opus 4.6 (Adaptive) | 9.4% | estimated ± 9.5 pp, medium confidence |
| 100 | Kimi K2.7 Code | 9.4% | estimated ± 9.5 pp, medium confidence |
| 101 | Step 3.7 Flash | 9.2% | estimated ± 8.3 pp, medium confidence |
| 102 | Claude Sonnet 4.6 | 9.0% | estimated ± 9.5 pp, low confidence |
| 103 | MiniMax M2.7 | 9.0% | estimated ± 8.3 pp, medium confidence |
| 104 | MiMo-V2.5 | 8.8% | estimated ± 8.3 pp, medium confidence |
| 105 | Claude Opus 4.5 Thinking | 8.6% | estimated ± 9.5 pp, low confidence |
| 106 | Gemini 3 Pro | 8.6% | estimated ± 9.5 pp, low confidence |
| 107 | Inkling-Small | 8.5% | estimated ± 8.3 pp, medium confidence |
| 108 | Gemini 3 Flash | 8.0% | estimated ± 9.5 pp, low confidence |
| 109 | GPT-5.2 | 8.0% | estimated ± 8.3 pp, medium confidence |
| 110 | DeepSeek V4 Pro | 7.6% | estimated ± 8.3 pp, medium confidence |
| 111 | GPT-5.4 mini | 7.3% | estimated ± 9.5 pp, low confidence |
| 112 | K-EXAONE 2.0 | 7.2% | estimated ± 14.2 pp, low confidence |
| 113 | Gemini 3.5 Flash | 7.1% | estimated ± 8.3 pp, medium confidence |
| 114 | GLM-5 | 7.1% | estimated ± 8.3 pp, medium confidence |
| 115 | GPT-5.2-Codex | 7.1% | estimated ± 9.5 pp, low confidence |
| 116 | Grok 4.3 | 6.8% | estimated ± 9.5 pp, low confidence |
| 117 | Claude Sonnet 4.5 Thinking | 6.4% | estimated ± 9.5 pp, low confidence |
| 118 | GPT-5.1 | 6.3% | estimated ± 9.5 pp, low confidence |
| 119 | GPT-5.4 nano | 6.3% | estimated ± 9.5 pp, low confidence |
| 120 | A.X K2 | 6.3% | estimated ± 14.2 pp, low confidence |
| 121 | GLM-4.7 | 6.2% | estimated ± 9.5 pp, low confidence |
| 122 | GPT-5 | 6.1% | estimated ± 9.5 pp, low confidence |
| 123 | Nemotron 3 Ultra | 6.1% | estimated ± 9.5 pp, low confidence |
| 124 | Trinity-Large-Preview | 5.9% | estimated ± 14.2 pp, low confidence |
| 125 | Trinity-Large-Thinking | 5.9% | estimated ± 14.2 pp, low confidence |
| 126 | Spark-X2.6-4B | 5.7% | estimated ± 8.3 pp, medium confidence |
| 127 | Gemini 3.5 Flash-Lite | 5.6% | estimated ± 8.3 pp, low confidence |
| 128 | Claude Haiku 4.5 | 5.5% | estimated ± 9.5 pp, low confidence |
| 129 | Mistral Medium 3.5 128B | 5.5% | estimated ± 9.5 pp, low confidence |
| 130 | Gemma 4 26B A4B | 5.5% | estimated ± 14.2 pp, low confidence |
| 131 | Nemotron 3.5 Bolt 74B-A9B | 5.2% | estimated ± 14.2 pp, low confidence |
| 132 | Qwen3.5-122B-A10B | 5.2% | estimated ± 14.2 pp, low confidence |
| 133 | Qwen3.5 Flash | 5.0% | estimated ± 9.5 pp, low confidence |
| 134 | Gemini 3.1 Flash-Lite | 4.7% | estimated ± 9.5 pp, low confidence |
| 135 | DeepSeek V3 0324 | 4.7% | estimated ± 14.2 pp, low confidence |
| 136 | GPT-OSS 20B | 4.6% | estimated ± 14.2 pp, low confidence |
| 137 | Mistral Small 4 | 4.6% | estimated ± 14.2 pp, low confidence |
| 138 | North Mini Code | 4.6% | estimated ± 14.2 pp, low confidence |
| 139 | Command A+ | 4.4% | estimated ± 14.2 pp, low confidence |
| 140 | Mercury 2.5 | 4.4% | estimated ± 14.2 pp, low confidence |
| 141 | Qwen3.6-27B | 4.4% | estimated ± 8.3 pp, low confidence |
| 142 | Claude Opus 4.6 | 4.2% | estimated ± 8.3 pp, low confidence |
| 143 | Inkling | 4.1% | measured |
| 144 | MiMo-V2-Flash | 3.6% | estimated ± 14.3 pp, low confidence |
| 145 | Mistral Large 3 | 3.3% | estimated ± 14.2 pp, low confidence |
| 146 | MAI-Thinking-1 | 3.1% | estimated ± 8.3 pp, low confidence |
| 147 | DeepSeek V3 | 2.9% | estimated ± 14.2 pp, low confidence |
| 148 | DeepSeek V4 Flash 0731 | 2.8% | estimated ± 8.3 pp, low confidence |
| 149 | Muse Spark | 2.5% | estimated ± 8.3 pp, low confidence |
| 150 | GPT-OSS 120B | 2.1% | estimated ± 14.2 pp, low confidence |
| 151 | Nemotron 3.5 Lightning 30B A3B NVFP4 | 1.5% | estimated ± 14.2 pp, low confidence |
| 152 | Grok 4.20 | 1.4% | estimated ± 8.3 pp, low confidence |
| 153 | Llama 4 Maverick | 1.4% | estimated ± 14.2 pp, low confidence |
| 154 | Nemotron 3.5 Super VL 120B A12B | 1.4% | estimated ± 8.3 pp, low confidence |
| 155 | o1 | 0.9% | estimated ± 14.3 pp, low confidence |
| 156 | Mistral Small 3.1 | 0.6% | estimated ± 14.2 pp, low confidence |
| 157 | Nemotron 3.5 Super VL 120B A12B NVFP4 | 0.5% | estimated ± 8.3 pp, low confidence |
| 158 | MiniCPM5-2B | 0.5% | estimated ± 14.2 pp, low confidence |
| 159 | o1-preview | 0.4% | estimated ± 14.3 pp, low confidence |
| 160 | Muse Glimmer 30B | 0.4% | estimated ± 8.3 pp, low confidence |
| 161 | K-Exaone | 0.3% | estimated ± 14.3 pp, low confidence |
| 162 | Composer 2.5 | 0.2% | estimated ± 5.8 pp, low confidence |
| 163 | Gemma 4 12B | 0.2% | estimated ± 14.3 pp, low confidence |
| 164 | Ling 2.6 Flash | 0.1% | estimated ± 14.3 pp, low confidence |
| 165 | Gemini 1.5 Pro | 0.0% | estimated ± 14.3 pp, low confidence |
| 166 | Granite 4.2 30B | 0.0% | estimated ± 8.3 pp, low confidence |
| 167 | Granite 4.2 8B | 0.0% | estimated ± 8.3 pp, low confidence |
| 168 | Kimi K2.5 | 0.0% | estimated ± 8.3 pp, low confidence |
| 169 | Laguna M.1 | 0.0% | estimated ± 8.3 pp, low confidence |
| 170 | Laguna XS.2 | 0.0% | estimated ± 8.3 pp, low confidence |
| 171 | Laguna XS 2.1 | 0.0% | estimated ± 8.3 pp, low confidence |
| 172 | LLaDA2.2-flash | 0.0% | estimated ± 8.3 pp, low confidence |
| 173 | LongCat-Flash-Lite-Sparse | 0.0% | estimated ± 8.3 pp, low confidence |
| 174 | Mellum2.1-12B-A2.5B-Thinking | 0.0% | estimated ± 8.3 pp, low confidence |
| 175 | Ornith-1.0-35B | 0.0% | estimated ± 8.3 pp, low confidence |
| 176 | Ornith-1.0-9B | 0.0% | estimated ± 8.3 pp, low confidence |
| 177 | Ornith-1.5-9B | 0.0% | estimated ± 8.3 pp, low confidence |
| 178 | Qwen3.5 397B | 0.0% | estimated ± 8.3 pp, low confidence |
| 179 | Qwen3.6-35B-A3B | 0.0% | estimated ± 8.3 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