- Agentic terminal coding Terminal-Bench 2.0 59.1
- Agentic coding SWE-bench Verified 79.6
- Agentic computer use OSWorld-Verified 72.5
- Mean of Agentic tool use τ2-bench Retail / Agentic tool use τ2-bench Telecom 94.8
- Scaled tool use MCP-Atlas 61.3
- Agentic search BrowseComp 74.7
- Multidisciplinary reasoning Humanity's Last Exam (HLE) (with tools) 49
- Agentic financial analysis Finance Agent v1.1 63.3
- Office tasks GDPval-AA Elo 1633
- Novel problem-solving ARC-AGI-2 58.3
- Graduate-level reasoning GPQA Diamond 89.9
- Visual reasoning MMMU-Pro (with tools) 75.6
- Multilingual Q&A MMMLU 89.3
Claude Sonnet 4.6
Claude Sonnet 4.6 ranks first on 5 of the 210 benchmarks it has scores on.
221
Benchmarks with a score
5
Ranked first, out of five or more models
28
In the top three
Anthropic Feb 2026
Claude Sonnet 4.6
Closed weights Mid-size
152.9
#50 of 271
Epoch Capabilities Index
- SWE-bench Verified 75.2% #14 /32
- Terminal-Bench 2.0 53.4% #15 /43
- MMMU-Pro 75.6% #32 /72
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Knowledge and language
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| BullshitBench | 94.5% | #2 of 107 | Official leaderboard | |
| Judgemark v2.1 | 94.1 | #1 of 42 | Official leaderboard | |
| BullshitBench V2 | 91% | #3 of 122 | Official leaderboard | |
| EQ-Bench 3 | 1,710 | #5 of 78 | Official leaderboard | |
| MMLU-Pro | 87.3% | #14 of 212 | Official leaderboard | |
| LMArena (Elo) | 1,472 | #39 of 333 | Official leaderboard | |
| Judgemark v4 | 82.2 | #5 of 40 | Official leaderboard | |
| Longform Creative Writing (EQ-Bench) | 79.9 | #15 of 119 | Official leaderboard | |
| HELM Arabic Enterprise | 73.7 | #4 of 23 | Official leaderboard | |
| Phare | 74% | #11 of 62 | Official leaderboard | |
| MMMLU | 89.3% | #12 of 62 | Model card | |
| Creative Writing v3 | 77.7 | #25 of 120 | Official leaderboard | |
| HELM Arabic | 81.5 | #6 of 28 | Official leaderboard | |
| Document Arena | 1,478 | #10 of 37 | Official leaderboard | |
| Humanity's Last Exam | 49% | #40 of 147 | Model card | |
| SEA-HELM | 74.1 | #18 of 62 | Official leaderboard | |
| FACTS Search | 76.4% | #9 of 31 | Official leaderboard | |
| Extended NYT Connections | 80.9% | #36 of 97 | Official leaderboard | |
| SimpleQA Verified | 30.9% | #30 of 56 | Official leaderboard | |
| EQ-Bench 4 | 1,197 | #20 of 32 | Official leaderboard | |
| Vectara Hallucination Leaderboard (2025 dataset) | 10.6% | #67 of 104 | Official leaderboard | |
| MILU | 89.6% | #7 of 10 | Model card | |
| Global MMLU | 88.7% | #8 of 11 | Model card | |
| FACTS Multimodal | 32.5% | #29 of 39 | Official leaderboard | |
| FACTS Benchmark Suite | 32.5% | #35 of 40 | Official leaderboard | |
| SkateBench v2 | 15.1% | #20 of 21 | Official leaderboard |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| AIME 2025 | 95.6% | #23 of 143 | Model card | |
| FrontierMath | 32.4% | #17 of 70 | Epoch AI | |
| ProofBench | 45% | #16 of 59 | Official leaderboard | |
| GPQA Diamond | 87.4% | #61 of 219 | Epoch AI | |
| DTBench | 83.1% | #51 of 178 | Official leaderboard | |
| ARC-AGI-2 | 60.4% | #34 of 95 | Official leaderboard | |
| OTIS Mock AIME 2024-2025 | 85.8% | #77 of 202 | Epoch AI | |
| FrontierMath Tier 4 | 8.30% | #22 of 55 | Epoch AI | |
| ARC-AGI | 86.5% | #42 of 92 | Official leaderboard | |
| Verifying the Verifiers (Project APE CRED benchmark) | 83.5% | #11 of 23 | Official leaderboard | |
| CritPt | 3.14% | #77 of 160 | Epoch AI | |
| ForecastBench | 64.0 | #61 of 86 | Official leaderboard | |
| BixBench3 | 15.3 | #18 of 20 | Official leaderboard | |
| LatchBio Capabilities v1.0 | 31.5% | #25 of 27 | Official leaderboard | |
| BioMysteryBench | 50.4% | #8 of 8 | Model card | |
| Organic Chemistry (Anthropic) | 48.4% | #5 of 5 | Model card | |
| BioPipelineBench (Anthropic) | 52.1% | #3 of 3 | Model card | |
| Phylogenetics (Anthropic) | 49.1% | #3 of 3 | Model card |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SWE-bench Verified (hard subset) | 48.2% | #1 of 9 | Model card | |
| Vibe Code Bench v1 | 29.1% | #3 of 17 | Official leaderboard | |
| WeirdML | 66.1% | #25 of 112 | Official leaderboard | |
| ALE-Bench | 1,327 | #29 of 121 | Epoch AI | |
| WebDev Arena | 1,521 | #37 of 105 | Epoch AI | |
| GBAEval | 48.8% | #8 of 21 | Official leaderboard | |
| SWE-bench Multilingual | 75.9% | #27 of 70 | Model card | |
| Code Arena | 1,523 | #48 of 124 | Official leaderboard | |
| Code Migration | 39.9% | #31 of 78 | Official leaderboard | |
| Text-based reinforcement learning task (Anthropic) | 0.93 | #3 of 7 | Model card | |
| SWE-bench Verified | 75.2% | #14 of 32 | Epoch AI | |
| Vibe Code Bench | 55.8% | #57 of 115 | Official leaderboard | |
| Time series forecasting task (Anthropic) | 5.70 | #5 of 9 | Model card | |
| FrontierCode | 15.1% | #7 of 12 | Official leaderboard | |
| AndroidBench | 67% | #19 of 32 | Official leaderboard | |
| Image-to-WebDev Arena | 1,529 | #33 of 55 | Official leaderboard | |
| ProgramBench | 0% | #15 of 25 | Official leaderboard | |
| SlopCodeBench | 16.8% | #9 of 14 | Official leaderboard | |
| PostTrainBench | 16.4% | #15 of 23 | Official leaderboard | |
| SWE Atlas | 31.2% | #15 of 22 | Official leaderboard | |
| SWE Atlas Test Writing | 31.8% | #15 of 22 | Official leaderboard | |
| OpenHands Index | 44.5% | #23 of 33 | Official leaderboard | |
| SWE Atlas Refactoring | 32.2% | #13 of 17 | Official leaderboard | |
| DeepSWE | 29.9% | #27 of 28 | Official leaderboard | |
| FrontierCode 1.1 | 24.3% | #8 of 8 | Official leaderboard | |
| OpenRCA | 27.9% | #2 of 2 | Model card |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Claw-Eval | 81.4% | #1 of 23 | Official leaderboard | |
| DeepResearch Bench | 54.9% | #2 of 24 | Epoch AI | |
| LisanBench | 12060.0 | #12 of 133 | Official leaderboard | |
| tau2-bench | 94.8% | #7 of 69 | Model card | |
| OSWorld | 72.1% | #1 of 9 | Epoch AI | |
| OSWorld-Verified | 81.5% | #4 of 33 | Official leaderboard | |
| Claw-Eval Pass^3 | 67.8% | #3 of 23 | Official leaderboard | |
| PinchBench | 76.4% | #9 of 48 | Official leaderboard | |
| UniClawBench | 45.5% | #2 of 10 | Official leaderboard | |
| Pencil Puzzle Bench | 26.7% | #11 of 53 | Official leaderboard | |
| WebArena | 65.6% | #2 of 8 | Model card | |
| Toolathlon | 44.8% | #13 of 50 | Official leaderboard | |
| SnakeBench | 26 | #55 of 183 | Official leaderboard | |
| Claw-Eval-Live | 61.9% | #4 of 12 | Official leaderboard | |
| GameWorld | 39.3% | #4 of 12 | Official leaderboard | |
| ClawBench | 56.9% | #2 of 6 | Official leaderboard | |
| Terminal-Bench 2.0 | 53.4% | #15 of 43 | Official leaderboard | |
| PACT | 1,546 | #10 of 26 | Official leaderboard | |
| DeepSearchQA | 89.2% | #9 of 22 | Model card | |
| LLM Chess | 190 | #54 of 124 | Official leaderboard | |
| LMArena Search Arena | 1,202 | #16 of 35 | Official leaderboard | |
| WebArena-Verified | 66.4% | #3 of 6 | Model card | |
| Long-Horizon-Terminal-Bench (LHTB) | 37.3% | #13 of 25 | Official leaderboard | |
| Game Arena Clobber | 265 | #19 of 36 | Official leaderboard | |
| BrowseComp | 74.7% | #41 of 77 | Model card | |
| GAIA | 54.5% | #31 of 57 | Official leaderboard | |
| Game Arena Word Association | 201 | #20 of 36 | Official leaderboard | |
| MyPCBench | 50.5% | #5 of 9 | Official leaderboard | |
| OSWorld 2.0 | 9.30% | #4 of 7 | Official leaderboard | |
| Game Arena Werewolf | -2 | #5 of 8 | Official leaderboard | |
| Game Arena Coin Game | 287 | #22 of 35 | Official leaderboard | |
| Chess Puzzles | 23% | #53 of 83 | Epoch AI | |
| Game Arena Dots and Boxes | 805 | #24 of 37 | Official leaderboard | |
| SkillsBench | 47.2% | #14 of 21 | Official leaderboard | |
| iOSWorld | 46.6% | #2 of 3 | Official leaderboard | |
| MCP Atlas | 69.5% | #23 of 34 | Official leaderboard | |
| Game Arena Checkers | 690 | #26 of 38 | Official leaderboard | |
| CEO-Bench | $38,166 | #13 of 19 | Official leaderboard | |
| Game Arena Ultimate Tic-Tac-Toe | 188 | #25 of 36 | Official leaderboard | |
| PinchBench 2.0 | 62.7% | #40 of 57 | Official leaderboard | |
| Kaggle Game Arena | 200 | #28 of 39 | Official leaderboard | |
| Mystery Game Puzzles | 16% | #60 of 80 | Epoch AI | |
| Game Arena Reversi | 228 | #27 of 36 | Official leaderboard | |
| CUA-World-Long | 6.00% | #3 of 4 | Official leaderboard | |
| Game Arena Word Art | 219 | #27 of 35 | Official leaderboard | |
| Game Arena Nine Men's Morris | 426 | #28 of 36 | Official leaderboard | |
| Game Arena Dark Hex | 115 | #31 of 38 | Official leaderboard | |
| Game Arena Lines of Action | 323 | #30 of 36 | Official leaderboard | |
| Game Arena Four in a Row | 136 | #34 of 38 | Official leaderboard | |
| Game Arena Five in a Row | 241 | #18 of 20 | Official leaderboard | |
| Game Arena Poker | -19.86 | #28 of 31 | Official leaderboard | |
| Game Arena Go | 296 | #33 of 36 | Official leaderboard | |
| Game Arena Chess | 98 | #36 of 39 | Official leaderboard | |
| Game Arena Bargaining | 78 | #35 of 36 | Official leaderboard | |
| ApprenticeBench | 2.00% | #18 of 18 | Official leaderboard |
Professional work
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| TaxEval v2 | 77.1% | #4 of 129 | Official leaderboard | |
| Vals Finance Agent | 63.3% | #2 of 49 | Official leaderboard | |
| ProfBench | 55.6% | #15 of 72 | Official leaderboard | |
| EnterpriseClawBench | 64.4% | #2 of 9 | Official leaderboard | |
| MortgageTax | 67.7% | #22 of 86 | Official leaderboard | |
| CorpFin v2 | 65.3% | #31 of 119 | Official leaderboard | |
| CaseLaw v2 | 64% | #14 of 53 | Official leaderboard | |
| Vending-Bench 2 | $7,204 | #20 of 66 | Official leaderboard | |
| GDPval-AA | 1,633 | #5 of 16 | Model card | |
| Legal Research Bench | 38.5% | #35 of 88 | Official leaderboard | |
| RCM-Bench-1 | 89.8% | #8 of 20 | Official leaderboard | |
| TaxCalcBench | 37.3% | #7 of 17 | Official leaderboard | |
| Excel Modeling Benchmark | 60.1% | #34 of 76 | Official leaderboard | |
| PhysicianBench | 23% | #6 of 13 | Official leaderboard | |
| Harvey LAB (Legal Agent Benchmark) | 5.00% | #37 of 80 | Official leaderboard | |
| Vending-Bench Arena | $5,639 | #15 of 32 | Official leaderboard | |
| Tax Agent Bench | 25.8% | #39 of 79 | Official leaderboard | |
| Vals Finance Agent v2 | 51% | #41 of 80 | Official leaderboard | |
| LegalBench | 82.1% | #73 of 138 | Official leaderboard | |
| Public Benefits Bench | 62.5% | #27 of 50 | Official leaderboard | |
| FORTE | 68.3% | #13 of 21 | Official leaderboard | |
| JobBench | 36.6% | #11 of 16 | Official leaderboard | |
| PolicyBench | 79.9% | #35 of 50 | Official leaderboard | |
| CADGenBench | 28.3% | #8 of 11 | Official leaderboard | |
| APEX-Agents | 43% | #41 of 56 | Epoch AI | |
| EEBench | 20.4% | #23 of 26 | Official leaderboard | |
| prinzbench | 8.00 | #26 of 26 | Official leaderboard | |
| Real-World Finance (Anthropic) | 59.5% | #3 of 3 | Model card | |
| MedCalc-Bench Verified | 86.2% | #1 of 1 | Model card |
Vision, audio and robotics
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Vision Arena | 1,275 | #27 of 136 | Official leaderboard | |
| Claw-Eval-MM | 50.9% | #4 of 11 | Official leaderboard | |
| ZeroBench | 12% | #26 of 66 | Official leaderboard | |
| MMMU-Pro | 75.6% | #32 of 72 | Model card | |
| GeoBench | 73% | #24 of 48 | Official leaderboard | |
| MDPBench | 73.1 | #5 of 9 | Official leaderboard | |
| GDP.pdf | 18% | #25 of 39 | Epoch AI | |
| 1H-VideoQA | 67.3% | #14 of 20 | Official leaderboard | |
| CharXiv | 72.4% | #47 of 67 | Model card | |
| Blueprint-Bench 2 | 6.70% | #29 of 35 | Official leaderboard |
Safety and security
Other
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| ACCoRD | 73.4 | #31 of 150 | Official leaderboard | |
| CAIS Risk Index | 40.4 | #14 of 65 | Official leaderboard | |
| LMCA | 46.5% | #34 of 139 | Epoch AI | |
| Conceptual Reasoning Index | 59.2 | #38 of 149 | Official leaderboard | |
| SAGE (Vals) | 46.6% | #34 of 81 | Official leaderboard | |
| Vals Multimodal Index | 60.6% | #17 of 33 | Official leaderboard | |
| CAIS Text Capabilities Index | 26.3 | #40 of 66 | Official leaderboard | |
| CAIS Vision Capabilities Index | 46.3 | #36 of 53 | Official leaderboard |
Anthropic's models over time
Anthropic's models in color, against every other model on the Epoch Capabilities Index, by release date.
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.
Its system card
Every result in the longer documents the lab published with it. 67 of them aren't in the launch post.
System Card: Claude Sonnet 4.6 17 Feb 2026
Capabilities
- SWE-bench Verified 79.6
- Terminal-Bench 2.0 59.1
- tau2-bench
- MCP Atlas 61.3
- OSWorld-Verified 72.5
- ARC-AGI-2 58.3
- GPQA Diamond 89.9
- MMMLU 89.3
- GDPval-AA 1633
- MMMU-Pro 74.5
- Humanity's Last Exam 33.2
- SWE-bench Multilingual 75.9
- OpenRCA 27.9
- ARC-AGI 86.5
- AIME 2025 95.6
- Vals Finance Agent 63.3
- Real-World Finance (Anthropic) 59.5
- Vending-Bench 2 7204.1
- CyberGym 65.2
- OpenAI MRCR v2
- GraphWalks
- LAB-Bench FigQA 58.8
- CharXiv 72.4
- WebArena 65.6
- WebArena-Verified 66.4
- Global MMLU 88.7
- MILU 89.6
- BrowseComp 74
- DeepSearchQA 89.2
- BioPipelineBench (Anthropic) 52.1
- BioMysteryBench 50.4
- Structural Biology (Anthropic)
- Organic Chemistry (Anthropic) 48.4
- Phylogenetics (Anthropic) 49.1
Health
Safety
- Single-turn harmful request evaluations (Anthropic) 99.4
- Single-turn benign request evaluations (Anthropic) 0.41
- Higher-difficulty harmful request evaluations (Anthropic) 99.4
- Higher-difficulty benign request evaluations (Anthropic) 0.18
- Multi-turn testing (Anthropic)
- Child safety evaluations (Anthropic)
- Suicide and self-harm evaluations (Anthropic)
- SSH stress-testing evaluation (Anthropic) 92
- Agentic coding malicious requests (Anthropic) 100
- Malicious use of Claude Code (Anthropic) 79.3
- Malicious computer use (Anthropic) 99.4
Robustness
Alignment
- Reward-hack-prone coding tasks (Anthropic) 0
- Impossible tasks (Anthropic) 40
- Agentic code behavior scores (Anthropic)
- Over-eager behavior on GUI computer use tasks (Anthropic) 14.7
- Anthropic automated behavioral audit
- Petri
- Refusal to assist with AI safety R&D (Anthropic) 38
- Self-preference bias (Anthropic)
- Participation in junk science (Anthropic) 0
- SHADE-Arena 1.6