Anthropic · model
Claude 3.5 Sonnet benchmark scores
As of 2026-11-21, Claude 3.5 Sonnet (Anthropic) has measured scores on 6 benchmarks and estimated scores on 30 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 63.3% | estimated ± 3.6 pp, high confidence |
| AA-HLE | 4.4% | estimated ± 5.7 pp, medium confidence |
| AA-SciCode | 32.7% | estimated ± 3.8 pp, high confidence |
| Artificial Analysis Intelligence Index | 7.5% | estimated ± 5.8 pp, medium confidence |
| CritPt | 0.3% | estimated ± 4.0 pp, high confidence |
| GPQA Diamond | 59.7% | estimated ± 1.2 pp, high confidence |
| HLE | 4.7% | estimated ± 6.5 pp, low confidence |
| AA-Omniscience Accuracy | 21.8% | estimated ± 6.5 pp, medium confidence |
| SciCode | 32.1% | estimated ± 3.6 pp, high confidence |
| AA Coding Index | 24.3% | estimated ± 5.3 pp, medium confidence |
| SWE-bench Verified | 49.0% | measured |
| SWE-bench Pro | 24.2% | estimated ± 4.5 pp, high confidence |
| MMLU-Pro | 75.9% | estimated ± 4.4 pp, high confidence |
| BBH | 82.7% | estimated ± 4.5 pp, low confidence |
| FailureSensorIQ | 58.6% | measured |
| GPQA | 59.4% | measured |
| MMLU | 88.2% | estimated ± 1.2 pp, high confidence |
| Vals GPQA Diamond | 59.3% | measured |
| Vals LiveCodeBench | 49.6% | measured |
| Vals MMLU-Pro | 78.4% | measured |
| MMMLU | 78.0% | estimated ± 1.7 pp, high confidence |
| AA MMLU-Pro | 83.3% | estimated ± 1.1 pp, medium confidence |
| ARC-AGI-1 | 58.1% | estimated ± 8.4 pp, medium confidence |
| CursorBench 3.2 | 59.1% | estimated ± 7.2 pp, low confidence |
| Vals SWE-bench | 34.5% | estimated ± 6.6 pp, low confidence |
| VulcanBench v3 | 76.3% | estimated ± 5.6 pp, low confidence |
| HLE w/o tools | 24.5% | estimated ± 4.9 pp, high confidence |
| BinaryAudit | 0.0% | estimated ± 8.9 pp, low confidence |
| OTelBench | 3.1% | estimated ± 7.7 pp, low confidence |
| Vibe Code Bench | 0.0% | estimated ± 8.4 pp, low confidence |
| C-Eval | 85.3% | estimated ± 1.2 pp, high confidence |
| MMLU-Redux | 87.3% | estimated ± 1.7 pp, high confidence |
| SuperGPQA | 41.3% | estimated ± 8.5 pp, low confidence |
| React Native Evals | 65.7% | estimated ± 2.7 pp, low confidence |
| SimpleQA | 14.6% | estimated ± 13.4 pp, low confidence |
| LiveCodeBench | 57.4% | estimated ± 14.6 pp, low confidence |