Anthropic · model
Claude Sonnet 4.5 benchmark scores
As of 2026-10-07, Claude Sonnet 4.5 (Anthropic) has measured scores on 11 benchmarks and estimated scores on 59 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 85.5% | estimated ± 2.7 pp, high confidence |
| AA-HLE | 30.9% | estimated ± 6.3 pp, medium confidence |
| AA-SciCode | 46.9% | estimated ± 3.9 pp, high confidence |
| AIME26 | 95.5% | estimated ± 0.6 pp, medium confidence |
| Artificial Analysis Intelligence Index | 18.1% | estimated ± 6.2 pp, medium confidence |
| CritPt | 1.1% | estimated ± 4.1 pp, high confidence |
| GDPval-AA | 32.8% | estimated ± 6.2 pp, medium confidence |
| GPQA Diamond | 83.4% | estimated ± 0.4 pp, high confidence |
| HLE | 43.8% | estimated ± 6.3 pp, medium confidence |
| LiveCodeBench v6 | 85.9% | estimated ± 8.1 pp, low confidence |
| AA-Omniscience Accuracy | 33.0% | estimated ± 6.2 pp, medium confidence |
| SciCode | 47.2% | estimated ± 3.8 pp, high confidence |
| BrowseComp | 65.1% | estimated ± 3.1 pp, high confidence |
| VITA-Bench | 17.0% | measured |
| AA Coding Index | 54.8% | estimated ± 5.1 pp, medium confidence |
| APEX-Agents-AA | 28.4% | estimated ± 3.6 pp, medium confidence |
| DeepSearchQA | 72.6% | estimated ± 8.5 pp, low confidence |
| SWE-bench Verified | 77.2% | measured |
| CyberGym | 43.3% | estimated ± 10.0 pp, low confidence |
| JobBench | 27.7% | measured |
| SWE-bench Pro | 53.4% | estimated ± 3.2 pp, high confidence |
| τ³-bench results | 68.9% | estimated ± 8.9 pp, medium confidence |
| WideResearch | 76.4% | estimated ± 1.2 pp, medium confidence |
| AA Agentic Index | 15.1% | estimated ± 6.4 pp, medium confidence |
| MMLU-Pro | 84.5% | estimated ± 3.0 pp, high confidence |
| FrontierMath v2 (Tier 4) | 4.2% | measured |
| FrontierMath v2 (Tiers 1-3) | 13.5% | measured |
| GPQA | 83.4% | measured |
| Gert Labs | 48.5% | measured |
| AA EnterpriseOps-Gym | 29.9% | estimated ± 5.9 pp, low confidence |
| ApprenticeBench | 0.0% | estimated ± 3.1 pp, low confidence |
| ARC-AGI-1 | 61.4% | estimated ± 1.9 pp, high confidence |
| ARC-AGI-2 | 13.6% | measured |
| FrontierCode 1.1 Main | 16.3% | estimated ± 2.7 pp, low confidence |
| OSWorld-Verified | 61.4% | measured |
| Vals GPQA Diamond | 80.7% | estimated ± 2.2 pp, high confidence |
| Vals LiveCodeBench | 79.1% | estimated ± 9.9 pp, low confidence |
| Vals MMLU-Pro | 78.8% | estimated ± 1.2 pp, high confidence |
| Vals SWE-bench | 72.5% | estimated ± 6.8 pp, medium confidence |
| Terminal-Bench 2.1 (Vals AI) | 43.7% | estimated ± 6.3 pp, medium confidence |
| AA AutomationBench | 47.3% | estimated ± 5.0 pp, low confidence |
| GDP.pdf | 17.8% | estimated ± 5.2 pp, low confidence |
| HLE w/o tools | 17.4% | estimated ± 4.9 pp, high confidence |
| OSWorld 2.0 | 0.0% | estimated ± 3.6 pp, medium confidence |
| Toolathlon-Verified | 73.2% | estimated ± 1.3 pp, low confidence |
| Vibe Code Bench | 14.7% | estimated ± 8.1 pp, medium confidence |
| C-Eval | 89.4% | estimated ± 0.9 pp, low confidence |
| Claw-Eval | 50.8% | estimated ± 5.6 pp, low confidence |
| DeepPlanning | 26.2% | estimated ± 9.1 pp, low confidence |
| HMMT Nov 2025 | 93.3% | estimated ± 2.6 pp, medium confidence |
| MCP Atlas | 51.2% | estimated ± 8.2 pp, medium confidence |
| MMAnswerBench | 82.9% | estimated ± 1.3 pp, medium confidence |
| MMLU-Redux | 93.0% | estimated ± 2.1 pp, high confidence |
| NL2Repo | 41.6% | estimated ± 6.1 pp, medium confidence |
| QwenClawBench | 52.0% | estimated ± 2.0 pp, low confidence |
| SuperGPQA | 65.3% | estimated ± 9.2 pp, low confidence |
| Terminal-Bench 2.0 | 50.0% | measured |
| Toolathlon | 40.8% | estimated ± 7.2 pp, medium confidence |
| HealthBench Hard | 0.0% | estimated ± 5.1 pp, low confidence |
| ResearchClawBench | 15.6% | estimated ± 2.2 pp, medium confidence |
| SWE-Rebench | 60.2% | estimated ± 3.6 pp, medium confidence |
| ARC-AGI-3 | 0.0% | estimated ± 2.1 pp, medium confidence |
| FrontierMath (legacy) | 14.5% | estimated ± 0.7 pp, low confidence |
| HealthBench Professional | 55.1% | estimated ± 4.6 pp, medium confidence |
| HLE-Verified | 35.3% | estimated ± 0.7 pp, low confidence |
| LABBench2 | 80.7% | estimated ± 2.5 pp, low confidence |
| AIME 2025 | 87.0% | measured |
| LiveCodeBench | 84.7% | estimated ± 10.2 pp, low confidence |
| Agents' Last Exam | 48.9% | estimated ± 8.7 pp, low confidence |
| MMLU | 90.2% | estimated ± 3.6 pp, low confidence |