| Benchmark | Claude Sonnet 5 | o4-mini |
|---|---|---|
| AA Agentic Index | 49.7 | 36.1 |
| AA Intelligence | 55.3 | 26.1 |
| AA-LCR | 77 | 60 |
| AA-Omniscience | 16.4 | -35.7 |
| Artificial Analysis Coding Index | 71.5 | 25.6 |
| browsecomp | 84.7 | 51.5 |
| critpt | 16.9 | 0.6 |
| gdpval | 54.8 | 25.4 |
| GPQA Diamond | 91.1 | 81.4 |
| HLE | 57.4 | 16.5 |
| MMMU-Pro | 77.3 | 69.2 |
| OmniScience Accuracy | 40 | 24.8 |
| OmniScience Non-Hallucination | 60.6 | 19.5 |
| scicode | 53.6 | 46.5 |
| simplebench | 60.6 | 38.7 |
| simpleqa_verified | 25 | 23.9 |
| SWE-bench Verified | 85.2 | 68.1 |
Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing.