| Benchmark | GPT-4o mini | Qwen3.7 Max |
|---|---|---|
| AA Agentic Index | 1 | 30.9 |
| AA Intelligence | 7 | 46.7 |
| Artificial Analysis Coding Index | 11.4 | 66 |
| gdpval | 0 | 38.5 |
| GPQA Diamond | 42.6 | 92.4 |
| HLE | 4.2 | 41.4 |
| IFBench | 31 | 80.5 |
| scicode | 22.9 | 53.5 |
| simplebench | 10.7 | 70.4 |
| SWE-bench Verified | 8.7 | 80.4 |
| TauBench V3 - Banking | 2.9 | 11.8 |
| Terminal-Bench 2.1 | 5.6 | 75 |
| τ³-Bench | 2.9 | 10.9 |
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.