- Agentic scientific research Terminal-Bench-Science 0.1 52.6
- Agentic coding Terminal-Bench 4.0 (Mythos 5.1) 60.9
- Knowledge work GDPval-AA v2 1853
- Computer use OSWorld 2.0 (partial) 77.9
- Multidisciplinary reasoning Humanity's Last Exam (with tools) 65
- Business workflows AutomationBench 31.4
- Agentic coding CursorBench 3.2.0 73.4
- SWE-bench Pro 81.2
- SWE-bench Multilingual 89.1
- SWE-bench Multimodal 54.7
- HealthBench Professional 62.1
- AA-Briefcase 1694
- ARC-AGI-1 97.5
- ARC-AGI-2 90
Claude Fable 5.1
Claude Fable 5.1 ranks first on 21 of the 210 benchmarks it has scores on.
215
Benchmarks with a score
21
Ranked first, out of five or more models
82
In the top three
Anthropic Sep 2026
Claude Fable 5.1
Closed weights Flagship
166.4
#6 of 271
Epoch Capabilities Index
- Humanity's Last Exam 46.5% #2 /47
- Terminal-Bench 4.0 57.9% #5 /23
- HealthBench Professional 62.1% #11 /25
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 | |
|---|---|---|---|---|
| LMArena (Elo) | 1,504 | #5 of 333 | Official leaderboard | |
| Longform Creative Writing (EQ-Bench) | 85.3 | #2 of 119 | Official leaderboard | |
| Creative Writing v3 | 81.3 | #5 of 120 | Official leaderboard | |
| Humanity's Last Exam | 46.5% | #2 of 47 | Official leaderboard | |
| BullshitBench V2 | 86.5% | #6 of 122 | Official leaderboard | |
| SimpleQA Verified | 70.8% | #6 of 84 | Epoch AI | |
| BullshitBench | 71.7% | #9 of 107 | Official leaderboard | |
| Document Arena | 1,493 | #4 of 37 | Official leaderboard | |
| Global MMLU | 94% | #2 of 11 | Model card | |
| MILU | 93% | #2 of 10 | Model card | |
| Extended NYT Connections | 90% | #22 of 97 | Official leaderboard | |
| HLE-Diamond | 55.1% | #5 of 18 | Official leaderboard |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| ProofBench | 100% | #1 of 83 | Epoch AI | |
| ProofBench v1.1 | 100% | #1 of 51 | Official leaderboard | |
| SimpleBench | 86.6% | #2 of 100 | Official leaderboard | |
| DTBench | 96% | #4 of 178 | Official leaderboard | |
| OTIS Mock AIME 2024-2025 | 100% | #5 of 202 | Epoch AI | |
| CritPt | 31.1% | #7 of 160 | Epoch AI | |
| FrontierMath Tiers 1-3 v2 | 90.2% | #5 of 89 | Epoch AI | |
| ARC-AGI-2 | 90% | #7 of 95 | Official leaderboard | |
| FrontierMath Tier 4 v2 | 87.8% | #7 of 71 | Epoch AI | |
| BixBench3 | 51.5 | #2 of 20 | Official leaderboard | |
| ARC-AGI | 97.5% | #10 of 92 | Official leaderboard | |
| LatchBio Capabilities v1.0 | 53% | #3 of 27 | Official leaderboard | |
| MysteryMechanism | 47.8% | #4 of 29 | Official leaderboard | |
| SciUniverse | 45.3% | #1 of 6 | Official leaderboard | |
| Terminal-Bench-Science | 40% | #3 of 17 | Official leaderboard | |
| PaperBenchX | 12.9% | #2 of 10 | Official leaderboard | |
| FrontierPhysics | 13.1% | #4 of 19 | Official leaderboard | |
| Prophet Arena | 85.6 | #5 of 18 | Official leaderboard | |
| Leipzig Benchmark | 79% | #7 of 19 | Official leaderboard | |
| CritPt v2 | 72% | #7 of 11 | Official leaderboard | |
| FrontierMath Erdős | 0% | #12 of 14 | Epoch AI | |
| ForecastBench | 66.0 | #84 of 86 | Official leaderboard |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| WeirdML | 92.9% | #3 of 112 | Official leaderboard | |
| GSO-Bench | 88.2% | #1 of 30 | Official leaderboard | |
| SWE-Bench Pro | 81.2% | #3 of 77 | Model card | |
| SWE Atlas Test Writing | 67% | #1 of 22 | Official leaderboard | |
| WebDev Arena | 1,758 | #5 of 105 | Epoch AI | |
| Code Arena | 1,746 | #6 of 124 | Official leaderboard | |
| Vibe Code Bench | 90.3% | #6 of 115 | Official leaderboard | |
| SWE-bench Multilingual | 89.1% | #4 of 70 | Model card | |
| PostTrainBench 1.2 | 44.6% | #1 of 17 | Official leaderboard | |
| ALE-Bench | 2,143 | #8 of 121 | Epoch AI | |
| MLS-Bench | 50.3% | #1 of 15 | Official leaderboard | |
| MirrorCode | 93.3% | #2 of 28 | Epoch AI | |
| Next.js Agent Evals | 97% | #1 of 14 | Official leaderboard | |
| Image-to-WebDev Arena | 1,722 | #4 of 55 | Official leaderboard | |
| ProgramBench | 9.00% | #2 of 25 | Official leaderboard | |
| Vals RSI Index | 0.36 | #2 of 25 | Official leaderboard | |
| SWE Atlas | 60% | #2 of 22 | Official leaderboard | |
| TasteVal | 171.6× | #2 of 20 | Official leaderboard | |
| AutoResearchExam | 0.60 | #1 of 9 | Official leaderboard | |
| SWE Atlas Refactoring | 56.7% | #2 of 17 | Official leaderboard | |
| SWE-Bench Pro V2 Hard | 92.2% | #2 of 16 | Official leaderboard | |
| FrontierCode | 50.9% | #6 of 46 | Epoch AI | |
| IOI v2 (Vals) | 90.8% | #6 of 45 | Official leaderboard | |
| Code Migration | 54.6% | #11 of 78 | Official leaderboard | |
| Vibe Code Bench 1-100 | 28% | #4 of 26 | Official leaderboard | |
| SWE-Bench Pro V2 | 99.1% | #3 of 16 | Official leaderboard | |
| FrontierSWE | 56.3% | #4 of 21 | Epoch AI | |
| WeirdML v3 | 26% | #5 of 25 | Official leaderboard | |
| CursorBench | 51.8% | #4 of 19 | Epoch AI | |
| RSI-Exam | 48.1% | #4 of 18 | Official leaderboard | |
| Senior SWE-Bench | 34.7% | #6 of 23 | Official leaderboard | |
| FrontierCode 2.0 | 32.4% | #7 of 26 | Official leaderboard | |
| DeepRSI | 1,548 | #2 of 7 | Official leaderboard | |
| Web Clone Bench | 13.3% | #3 of 9 | Official leaderboard | |
| FrontierCS 2 | 37.5% | #6 of 16 | Official leaderboard | |
| SWE-bench Multimodal | 54.7% | #4 of 10 | Model card | |
| SWE-sweep | 2.90% | #4 of 10 | Official leaderboard | |
| DeepSWE | 67.4% | #23 of 51 | Model card | |
| Android Bench 2.0 | 22.7% | #5 of 11 | Official leaderboard | |
| SWE-Marathon v1.1 | 45.6% | #8 of 17 | Official leaderboard | |
| FrontierCode 1.1 | 50.9% | #6 of 12 | Model card | |
| CursorBench 3.2 | 73.4% | #1 of 2 | Model card | |
| OpenDesign Arena | 80.3 | #13 of 19 | Official leaderboard |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| LisanBench | 15100.3 | #6 of 133 | Official leaderboard | |
| OSWorld 2.0 | 77.9% | #1 of 19 | Model card | |
| Game Arena Nine Men's Morris | 1,017 | #2 of 36 | Official leaderboard | |
| ApprenticeBench | 72% | #1 of 18 | Official leaderboard | |
| MCP Atlas | 87.2% | #2 of 34 | Official leaderboard | |
| Agent Arena | 0.13 | #4 of 53 | Official leaderboard | |
| Game Arena Go | 1,615 | #3 of 36 | Official leaderboard | |
| Mystery Game Puzzles | 58% | #9 of 80 | Epoch AI | |
| LMArena Search Arena | 1,233 | #4 of 35 | Official leaderboard | |
| Game Arena Poker | 17.9 | #4 of 31 | Official leaderboard | |
| Game Arena Four in a Row | 597 | #5 of 38 | Official leaderboard | |
| Game Arena Dots and Boxes | 1,116 | #5 of 37 | Official leaderboard | |
| Game Arena Clobber | 517 | #5 of 36 | Official leaderboard | |
| Game Arena Lines of Action | 774 | #5 of 36 | Official leaderboard | |
| Kaggle Game Arena | 418 | #6 of 39 | Official leaderboard | |
| Toolathlon-Verified | 77.8% | #2 of 13 | Model card | |
| EBR-bench | 57.1% | #5 of 28 | Epoch AI | |
| DRACO | 87.7% | #2 of 11 | Model card | |
| Game Arena Bargaining | 230 | #7 of 36 | Official leaderboard | |
| Terminal-Bench 4.0 | 57.9% | #5 of 23 | Official leaderboard | |
| Time Horizon Index: KSP | 63.3% | #3 of 13 | Official leaderboard | |
| Game Arena Checkers | 1,079 | #9 of 38 | Official leaderboard | |
| Chess Puzzles | 48% | #20 of 83 | Epoch AI | |
| Game Arena Reversi | 740 | #9 of 36 | Official leaderboard | |
| Game Arena Ultimate Tic-Tac-Toe | 442 | #9 of 36 | Official leaderboard | |
| CUA-bench | 13.2% | #3 of 9 | Official leaderboard | |
| Game Arena Dark Hex | 257 | #18 of 38 | Official leaderboard | |
| Game Arena Chess | 760 | #20 of 39 | Official leaderboard | |
| StarSkirmish Bench | 37.2% | #5 of 9 | Official leaderboard | |
| SpireClimb StS1 A20 Heart | 0% | #4 of 7 | Epoch AI | |
| AutomationBench | 31.4% | #26 of 41 | Model card | |
| SpireClimb | 2.00% | #5 of 7 | Epoch AI | |
| SpireClimb StS2 A10 | 4.00% | #5 of 7 | Epoch AI | |
| OSWorld 2.0 (Aug 2026) | 77.9% | #1 of 1 | Model card |
Professional work
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Excel Modeling Benchmark | 76.7% | #1 of 76 | Official leaderboard | |
| LegalBench | 88.5% | #2 of 138 | Official leaderboard | |
| MedScribe | 91.3% | #2 of 98 | Official leaderboard | |
| MortgageTax | 70.8% | #2 of 86 | Official leaderboard | |
| Excel Modeling Benchmark v1.1 | 76.7% | #1 of 37 | Official leaderboard | |
| Legal Research Bench | 55.3% | #3 of 88 | Official leaderboard | |
| Tax Agent Bench | 49.2% | #3 of 79 | Official leaderboard | |
| Public Benefits Bench | 74.9% | #2 of 50 | Official leaderboard | |
| Immigration Bench | 75.4% | #1 of 24 | Official leaderboard | |
| FrontierFinance | 55.9% | #1 of 16 | Official leaderboard | |
| TaxEval v2 | 76% | #9 of 129 | Official leaderboard | |
| GDPval-AA v2 | 1,853 | #2 of 26 | Model card | |
| $OneMillion-Bench | 42.1% | #1 of 13 | Official leaderboard | |
| RealmPathology | 78% | #1 of 13 | Official leaderboard | |
| MedCode | 53.5% | #8 of 96 | Official leaderboard | |
| APEX-Agents | 68.6% | #5 of 56 | Epoch AI | |
| Vals Finance Agent v2 | 58.9% | #8 of 80 | Official leaderboard | |
| Holdings-IndividualTaxBench | 75.5% | #2 of 13 | Official leaderboard | |
| PolicyBench | 90.8% | #8 of 50 | Official leaderboard | |
| OfficeQA | 80.2% | #2 of 12 | Model card | |
| Epoch Automation Reports | 65.3% | #1 of 6 | Official leaderboard | |
| Legora BAR | 104.0 | #3 of 17 | Official leaderboard | |
| Remote Labor Index | 17.9% | #3 of 17 | Official leaderboard | |
| PRBench | 51.6% | #8 of 44 | Official leaderboard | |
| Thrive Holdings helpdesk ticket resolution benchmark | 55.8% | #3 of 15 | Official leaderboard | |
| Build-A-Chip-Bench | 16.6% | #1 of 5 | Official leaderboard | |
| OfficeQA Pro | 69% | #3 of 14 | Model card | |
| LegalReasonBench | 50.6% | #2 of 8 | Official leaderboard | |
| Clipping Bench | $1,068 | #5 of 19 | Official leaderboard | |
| Parametric CAD Bench | 56.8% | #3 of 11 | Official leaderboard | |
| FoodTruck Bench | $62,732 | #3 of 10 | Official leaderboard | |
| CAD Arena | 0.66 | #4 of 12 | Official leaderboard | |
| AA-Briefcase | 1,694 | #2 of 6 | Model card | |
| TaxCalcBench v2 | 54% | #9 of 26 | Official leaderboard | |
| HealthBench | 60% | #12 of 30 | Model card | |
| Harvey LAB (Legal Agent Benchmark) | 6.67% | #33 of 80 | Official leaderboard | |
| PharmaBench (micro1) | 12.8% | #5 of 12 | Official leaderboard | |
| EEBench | 56.4% | #11 of 26 | Official leaderboard | |
| Vending-Bench Arena | $5,719 | #14 of 32 | Official leaderboard | |
| HealthBench Professional | 62.1% | #11 of 25 | Model card | |
| MissionReport | 63.3% | #4 of 9 | Official leaderboard | |
| Vending-Bench 2 | $5,422 | #30 of 66 | Official leaderboard | |
| RWE-bench | 7.00% | #5 of 11 | Official leaderboard | |
| micro1 Medical Index | 68.2% | #5 of 8 | Official leaderboard | |
| BenchCAD | 43.7% | #8 of 10 | Model card |
Vision, audio and robotics
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| ParseBench | 78.1% | #2 of 42 | Official leaderboard | |
| ArtifactArena | 1,310 | #2 of 21 | Official leaderboard | |
| Vision Arena | 1,288 | #14 of 136 | Official leaderboard | |
| Drone-Bench | 91.2% | #3 of 19 | Official leaderboard | |
| Blueprint-Bench 2 | 41.9% | #6 of 35 | Official leaderboard | |
| Furniture Assembly Benchmark | 70% | #7 of 35 | Epoch AI | |
| GDP.pdf | 29.6% | #8 of 39 | Epoch AI | |
| GeoBench | 84% | #10 of 48 | Official leaderboard | |
| ClockBench | 59.4% | #8 of 37 | Official leaderboard | |
| Humanity's Sixth Sense | 40.8% | #7 of 29 | Official leaderboard | |
| ScienceVision | 0.46 | #3 of 11 | Official leaderboard | |
| Code2Video Benchmark | 1,536 | #8 of 19 | Official leaderboard | |
| Dual-Arm Sim | 37.8 | #5 of 10 | Official leaderboard | |
| Robo-bench | 5.83% | #5 of 7 | Official leaderboard | |
| Chartography | 42.6% | #11 of 12 | Model card |
Safety and security
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Single-turn benign request evaluations (Anthropic) | 0% | #1 of 18 | Model card | |
| OncologyBench | 44% | #2 of 30 | Official leaderboard | |
| TxBench-Oligo-Discovery | 54.9% | #2 of 28 | Official leaderboard | |
| EpiBench | 31.8% | #2 of 27 | Official leaderboard | |
| MetagenomicsBench | 61.7% | #2 of 26 | Official leaderboard | |
| TxBench-Antibody-Discovery | 63.3% | #3 of 28 | Official leaderboard | |
| BioSecBench-Function | 49% | #3 of 27 | Official leaderboard | |
| PCT (Polarized Contrastive Pairs) | 33.4% | #6 of 46 | Official leaderboard | |
| Prompt injection robustness against adaptive attackers, computer use (Anthropic) | 0.07% | #2 of 13 | Model card | |
| FORTRESS | 13.7% | #11 of 66 | Official leaderboard | |
| BBQ (Bias Benchmark for QA) | 89.9% | #5 of 24 | Model card | |
| Anthropic political even-handedness eval | 98.3% | #3 of 13 | Model card | |
| SRE-Bench | 22.9% | #8 of 33 | Official leaderboard | |
| Cyber harm coverage (Anthropic) | 100% | #1 of 4 | Model card | |
| BioSecBench-Surveillance | 49% | #7 of 27 | Official leaderboard | |
| BrowseSafe++ | 89.2% | #4 of 15 | Official leaderboard | |
| TxBench-Preclinical-Pharmacology | 59.7% | #8 of 28 | Official leaderboard | |
| KORA | 57.5% | #13 of 42 | Official leaderboard | |
| Malicious use of Claude Code (Anthropic) | 90.3% | #5 of 16 | Model card | |
| scBench-Long | 25.6% | #9 of 27 | Official leaderboard | |
| VariantBench | 40.7% | #9 of 27 | Official leaderboard | |
| CheatBench | 45.8% | #4 of 10 | Official leaderboard | |
| Rewind-attacker cyber adversarial robustness (Anthropic) | 4.50% | #2 of 5 | Model card | |
| CyberBench v1.1 (Vals) | 70.4% | #19 of 43 | Official leaderboard | |
| SpatialBench-Long | 29.8% | #13 of 27 | Official leaderboard | |
| DistressBench | 71.5% | #21 of 42 | Official leaderboard | |
| CWE-bench | 58% | #6 of 12 | Official leaderboard | |
| Malicious computer use (Anthropic) | 85.7% | #8 of 13 | Model card | |
| Prompt injection robustness against adaptive attackers, browser use (Anthropic) | 2.64% | #9 of 13 | Model card | |
| WellnessBench | 51% | #9 of 12 | Official leaderboard | |
| BioSecBench-Refusal | 5.60% | #17 of 22 | Official leaderboard | |
| Gray Swan IPI Arena benchmark | 0.10% | #8 of 10 | Model card | |
| Prompt injection robustness against adaptive attackers, coding (Anthropic) | 9.26% | #12 of 13 | Model card | |
| Single-turn harmful request evaluations (Anthropic) | 94.7% | #17 of 18 | Model card | |
| Secure and defensive coding (Anthropic) | 1.03% | #2 of 2 | Model card | |
| Gray Swan Shade knowledge elicitation through chat | 1.00% | #1 of 1 | Model card |
Other
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Conceptual Reasoning Index | 74.7 | #3 of 149 | Official leaderboard | |
| LMCA | 65.5% | #3 of 139 | Epoch AI | |
| LiveBench | 85.7% | #1 of 28 | Official leaderboard | |
| CAIS Risk Index | 30.3 | #3 of 65 | Official leaderboard | |
| ACCoRD | 81.0 | #8 of 150 | Official leaderboard | |
| CAIS Text Capabilities Index | 55.6 | #5 of 66 | Official leaderboard | |
| CAIS Vision Capabilities Index | 68.9 | #7 of 53 | Official leaderboard | |
| SAGE (Vals) | 48.5% | #27 of 81 | 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. 39 of them aren't in the launch post.
Capabilities
- SWE-Bench Pro 81.2
- SWE-bench Multilingual 89.1
- SWE-bench Multimodal 54.7
- DeepSWE 67.4
- FrontierCode 1.1 50.9
- FrontierSWE 57
- Terminal-Bench 4.0 55.8
- Terminal-Bench-Science 52.6
- CursorBench 3.2 73.4
- CritPt 88.4
- ProgramBench 87.6
- Humanity's Last Exam 60.9
- DRACO 87.7
- Chartography 42.6
- BenchCAD 43.7
- OSWorld 2.0 (Aug 2026) 77.9
- GDP.pdf 85.4
- OfficeQA 80.2
- OfficeQA Pro 69
- Harvey LAB (Legal Agent Benchmark) 19.1
- GDPval-AA v2 1853
- AA-Briefcase 1694
- Toolathlon-Verified 77.8
- AutomationBench 31.4
- ARC-AGI 97.5
- ARC-AGI-2 90
- Global MMLU 94
- MILU 93
Health
- HealthBench 60
- HealthBench Professional 62.1
Safety
- Single-turn harmful request evaluations (Anthropic) 94.7
- Single-turn benign request evaluations (Anthropic) 0
- Multi-turn testing (Anthropic)
- Child safety evaluations (Anthropic)
- Suicide and self-harm evaluations (Anthropic)
- Disordered eating evaluations (Anthropic)
- Election integrity evaluations (Anthropic)
- Malicious use of Claude Code (Anthropic) 90.3
- Malicious computer use (Anthropic) 85.7
Robustness
- Rewind-attacker cyber adversarial robustness (Anthropic) 4.5
- Gray Swan Shade knowledge elicitation through chat 1
- Gray Swan IPI Arena benchmark 0.1
- Prompt injection robustness against adaptive attackers, coding (Anthropic) 9.26
- Prompt injection robustness against adaptive attackers, computer use (Anthropic) 0.07
- Prompt injection robustness against adaptive attackers, browser use (Anthropic) 2.64