| Benchmark | Claude Sonnet 5 | GPT-4o mini |
|---|---|---|
| AA Agentic Index | 49.7 | 1 |
| AA Intelligence | 55.3 | 7 |
| Artificial Analysis Coding Index | 71.5 | 11.4 |
| gdpval | 54.8 | 0 |
| GPQA Diamond | 91.1 | 42.6 |
| HLE | 57.4 | 4.2 |
| MMMU-Pro | 77.3 | 41.5 |
| scicode | 53.6 | 22.9 |
| simplebench | 60.6 | 10.7 |
| SWE-bench Verified | 85.2 | 8.7 |
| TauBench V3 - Banking | 37.3 | 2.9 |
| Terminal-Bench 2.1 | 80.5 | 5.6 |
| τ³-Bench | 28.2 | 2.9 |
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.