Anthropic · model
Claude Opus 5 benchmark scores
As of 2026-10-07, Claude Opus 5 (Anthropic) has measured scores on 54 benchmarks and estimated scores on 62 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 93.2% | measured |
| AA-HLE | 54.9% | measured |
| AA-SciCode | 56.4% | measured |
| Artificial Analysis Intelligence Index | 50.8% | measured |
| CritPt | 29.1% | measured |
| GDPval-AA | 61.2% | measured |
| GPQA Diamond | 94.0% | estimated ± 1.0 pp, medium confidence |
| HLE | 64.7% | measured |
| AA-LCR | 79.3% | measured |
| LiveCodeBench v6 | 95.8% | estimated ± 2.1 pp, high confidence |
| AA-Omniscience Accuracy | 60.9% | measured |
| SciCode | 54.8% | estimated ± 2.9 pp, low confidence |
| Terminal-Bench 2.1 | 91.8% | estimated ± 6.6 pp, low confidence |
| BrowseComp | 90.8% | measured |
| HLE w/ tools | 64.7% | measured |
| LongBench v2 | 61.5% | estimated ± 5.9 pp, low confidence |
| VITA-Bench | 50.4% | estimated ± 9.9 pp, low confidence |
| τ²-bench results | 94.2% | estimated ± 7.1 pp, low confidence |
| AA Coding Index | 78.0% | measured |
| AA-MMMU-Pro | 84.7% | measured |
| APEX-Agents | 38.5% | estimated ± 0.6 pp, low confidence |
| APEX-Agents-AA | 38.9% | estimated ± 2.9 pp, low confidence |
| BioMysteryBench (human-difficult) | 49.4% | measured |
| DeepSearchQA | 95.0% | measured |
| SWE-bench Verified | 96.0% | measured |
| AutomationBench | 26.0% | measured |
| BFCL v4 | 78.2% | estimated ± 2.6 pp, medium confidence |
| CyberGym | 89.1% | estimated ± 2.4 pp, low confidence |
| JobBench | 69.8% | estimated ± 3.4 pp, low confidence |
| SWE-bench Pro | 79.2% | measured |
| τ³-bench results | 69.6% | estimated ± 1.8 pp, low confidence |
| WideResearch | 86.5% | estimated ± 3.4 pp, low confidence |
| AA Agentic Index | 56.2% | measured |
| MMLU-Pro | 86.0% | estimated ± 2.5 pp, low confidence |
| GPQA | 95.6% | estimated ± 0.8 pp, high confidence |
| Gert Labs | 81.4% | estimated ± 3.0 pp, low confidence |
| AA-AnalystAgent | 53.8% | measured |
| AA EnterpriseOps-Gym | 54.3% | estimated ± 1.8 pp, low confidence |
| AA Harvey LAB | 93.5% | measured |
| ApprenticeBench | 36.0% | measured |
| ARC-AGI-1 | 97.5% | measured |
| ARC-AGI-2 | 90.4% | measured |
| CursorBench 3.1 | 59.8% | estimated ± 3.4 pp, medium confidence |
| CursorBench 3.2 | 70.0% | measured |
| Terminal-Bench 3.0 | 42.7% | measured |
| FrontierCode 1.1 Main | 53.4% | measured |
| FrontierSWE v2 | 52.0% | measured |
| OfficeQA Pro | 66.9% | measured |
| OSWorld-Verified | 86.4% | estimated ± 1.9 pp, low confidence |
| Vals GPQA Diamond | 93.4% | measured |
| Vals LiveCodeBench | 89.0% | measured |
| Vals MMLU-Pro | 91.6% | measured |
| Vals SWE-bench | 97.0% | measured |
| Terminal-Bench 2.1 (Vals AI) | 84.6% | measured |
| VulcanBench v3 | 87.0% | measured |
| AA AutomationBench | 61.3% | estimated ± 5.7 pp, low confidence |
| GDP.pdf | 27.1% | estimated ± 3.8 pp, high confidence |
| AA ITBench | 43.2% | estimated ± 3.4 pp, medium confidence |
| AA Tau3 Banking | 45.9% | estimated ± 4.4 pp, medium confidence |
| AA Terminal-Bench 2.1 | 88.8% | estimated ± 4.4 pp, high confidence |
| AA Terminal-Bench 4.0 | 41.5% | estimated ± 14.0 pp, low confidence |
| CursorBench 4.0 | 46.6% | measured |
| CWE-bench v1 | 62.7% | estimated ± 6.9 pp, medium confidence |
| DeepSWE | 68.8% | measured |
| HLE w/o tools | 56.3% | measured |
| OSWorld 2.0 | 70.6% | measured |
| PostTrainBench v1.1 | 35.0% | measured |
| ProgramBench | 93.0% | measured |
| SWE Multilingual | 89.5% | measured |
| Toolathlon-Verified | 80.6% | measured |
| Vibe Code Bench | 80.9% | estimated ± 5.6 pp, low confidence |
| CharXiv | 86.0% | estimated ± 7.5 pp, low confidence |
| CharXiv w/o tools | 85.5% | estimated ± 2.4 pp, high confidence |
| ExploitGym | 34.4% | estimated ± 6.1 pp, low confidence |
| C-Eval | 93.3% | estimated ± 0.8 pp, low confidence |
| Claw-Eval | 73.1% | estimated ± 5.7 pp, low confidence |
| DeepPlanning | 52.7% | estimated ± 13.1 pp, low confidence |
| MathVision | 89.7% | estimated ± 4.9 pp, medium confidence |
| MCP Atlas | 85.8% | measured |
| MCP-Tasks | 77.1% | estimated ± 0.6 pp, low confidence |
| MMLU-ProX | 90.1% | estimated ± 1.7 pp, low confidence |
| MMLU-Redux | 97.9% | estimated ± 1.1 pp, low confidence |
| MMMU-Pro | 81.3% | estimated ± 4.4 pp, low confidence |
| NL2Repo | 60.9% | estimated ± 4.7 pp, medium confidence |
| NOVA-63 | 62.9% | estimated ± 1.6 pp, low confidence |
| QwenClawBench | 70.9% | estimated ± 2.6 pp, medium confidence |
| ScreenSpot Pro | 84.4% | estimated ± 10.7 pp, low confidence |
| SuperGPQA | 89.3% | estimated ± 7.5 pp, low confidence |
| Terminal-Bench 2.0 | 82.5% | estimated ± 5.9 pp, low confidence |
| Toolathlon | 59.3% | estimated ± 1.7 pp, low confidence |
| ERQA | 68.9% | estimated ± 3.9 pp, medium confidence |
| HealthBench Hard | 34.9% | estimated ± 7.7 pp, low confidence |
| LiveCodeBench Pro | 96.8% | estimated ± 6.0 pp, low confidence |
| MedXpertQA (MM) | 77.8% | estimated ± 6.5 pp, low confidence |
| MedXpertQA (Text) | 72.4% | estimated ± 8.9 pp, low confidence |
| MMLU-Pro (Arcee) | 100.0% | estimated ± 4.0 pp, low confidence |
| React Native Evals | 100.0% | estimated ± 2.4 pp, low confidence |
| ResearchClawBench | 19.7% | estimated ± 1.8 pp, low confidence |
| SWE-Rebench | 61.8% | estimated ± 3.6 pp, low confidence |
| ARC-AGI-3 | 30.2% | measured |
| BioMysteryBench (human-solvable) | 90.1% | measured |
| DRACO | 88.6% | measured |
| FrontierCode 1.1 Extended | 63.6% | measured |
| HealthBench (raw) | 67.1% | measured |
| HealthBench (length-adjusted) | 57.8% | measured |
| HealthBench Professional | 59.8% | measured |
| HealthBench Professional (raw) | 73.4% | measured |
| HLE-Verified | 54.4% | measured |
| LABBench2 | 84.2% | measured |
| MMMU | 91.4% | estimated ± 0.9 pp, medium confidence |
| LiveCodeBench | 100.0% | estimated ± 8.9 pp, low confidence |
| OpenHarmony Bench | 58.8% | estimated ± 2.4 pp, high confidence |
| MMLU | 96.1% | estimated ± 1.1 pp, low confidence |
| MMMU-Pro w/ Python | 84.5% | estimated ± 1.6 pp, medium confidence |
| RealWorldQA | 90.0% | estimated ± 1.2 pp, low confidence |
| OmniDocBench 1.5 | 92.0% | estimated ± 6.4 pp, low confidence |