Anthropic · model
Claude 3.7 Sonnet Thinking benchmark scores
As of 2026-10-29, Claude 3.7 Sonnet Thinking (Anthropic) has measured scores on 13 benchmarks and estimated scores on 44 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 77.2% | measured |
| AA-HLE | 9.7% | measured |
| CritPt | 0.9% | measured |
| GDPval-AA | 1.9% | estimated ± 10.9 pp, low confidence |
| GPQA Diamond | 78.4% | estimated ± 1.2 pp, high confidence |
| HLE | 15.0% | estimated ± 6.9 pp, medium confidence |
| IFBench | 55.5% | estimated ± 6.6 pp, medium confidence |
| IFEval | 93.2% | measured |
| τ²-bench results | 54.7% | measured |
| AA-MMMU-Pro | 61.4% | estimated ± 3.0 pp, high confidence |
| BioMysteryBench (human-difficult) | 15.7% | estimated ± 6.8 pp, low confidence |
| BFCL v4 | 46.5% | estimated ± 8.5 pp, low confidence |
| AA Agentic Index | 10.7% | estimated ± 12.3 pp, low confidence |
| MMLU-Pro | 85.7% | estimated ± 4.1 pp, high confidence |
| AA-IFBench | 48.3% | measured |
| AIME 2024 | 61.3% | measured |
| Artificial Analysis Intelligence Index | 17.7% | measured |
| GPQA | 78.2% | measured |
| MMLU | 91.1% | estimated ± 0.7 pp, medium confidence |
| AA-Omniscience Accuracy | 28.0% | measured |
| MATH-500 | 96.2% | measured |
| MMMLU | 86.1% | measured |
| MMMU | 75.0% | measured |
| AA MMLU-Pro | 85.2% | estimated ± 0.6 pp, high confidence |
| AIME 2025 | 42.8% | estimated ± 3.3 pp, high confidence |
| ARC-AGI-1 | 42.2% | estimated ± 7.0 pp, low confidence |
| ARC-AGI-2 | 0.0% | estimated ± 11.6 pp, low confidence |
| Vals GPQA Diamond | 67.6% | estimated ± 4.8 pp, medium confidence |
| Vals MMLU-Pro | 62.4% | estimated ± 1.2 pp, medium confidence |
| HLE w/o tools | 31.5% | estimated ± 4.6 pp, high confidence |
| ARC-AGI-3 | 1.4% | estimated ± 7.7 pp, low confidence |
| AI2D_TEST | 89.5% | estimated ± 2.0 pp, high confidence |
| C-Eval | 90.9% | estimated ± 1.0 pp, medium confidence |
| CountBench | 94.6% | estimated ± 3.2 pp, high confidence |
| DynaMath | 82.8% | estimated ± 5.8 pp, low confidence |
| ERQA | 47.1% | estimated ± 5.2 pp, low confidence |
| MathVision | 74.0% | estimated ± 3.6 pp, high confidence |
| MLVU (M-Avg) | 83.3% | estimated ± 2.4 pp, high confidence |
| MMLU-Redux | 93.6% | estimated ± 0.5 pp, medium confidence |
| MMMU-Pro | 63.8% | estimated ± 2.3 pp, high confidence |
| RealWorldQA | 80.6% | estimated ± 3.6 pp, high confidence |
| ScreenSpot Pro | 53.0% | estimated ± 7.3 pp, low confidence |
| SimpleVQA | 56.1% | estimated ± 7.1 pp, medium confidence |
| SuperGPQA | 59.2% | estimated ± 8.2 pp, low confidence |
| Video-MME (w/o subtitle) | 75.9% | estimated ± 1.3 pp, high confidence |
| Video-MME (with subtitle) | 81.0% | estimated ± 3.2 pp, high confidence |
| VideoMMMU | 70.2% | estimated ± 3.8 pp, high confidence |
| We-Math | 77.4% | estimated ± 7.2 pp, low confidence |
| HealthBench Hard | 8.9% | estimated ± 8.6 pp, low confidence |
| MMLU-Pro (Arcee) | 80.5% | estimated ± 4.0 pp, medium confidence |
| BioMysteryBench (human-solvable) | 81.7% | estimated ± 0.6 pp, low confidence |
| HealthBench (length-adjusted) | 53.9% | estimated ± 3.3 pp, medium confidence |
| HealthBench Professional (raw) | 54.3% | estimated ± 8.6 pp, low confidence |
| HLE-Verified | 22.6% | estimated ± 7.1 pp, low confidence |
| STEMBench | 2.4% | estimated ± 0.2 pp, medium confidence |
| SimpleQA | 29.1% | estimated ± 4.3 pp, medium confidence |
| MMVU | 69.7% | estimated ± 3.8 pp, medium confidence |