| Benchmark | GPT-4o mini | Qwen3.5 397B A17B |
|---|---|---|
| AA Agentic Index | 1 | 53.3 |
| AA Intelligence | 7 | 34.3 |
| Artificial Analysis Coding Index | 11.4 | 48.2 |
| gdpval | 0 | 35.8 |
| GPQA Diamond | 42.6 | 89.3 |
| GSM8K | 84.3 | 66.7 |
| HLE | 4.2 | 29 |
| IFBench | 31 | 78.8 |
| MMMU | 59.4 | 85 |
| MMMU-Pro | 41.5 | 79 |
| scicode | 22.9 | 42 |
| SWE-bench Verified | 8.7 | 76.4 |
| TauBench V3 - Banking | 2.9 | 13.4 |
| Terminal-Bench 2.1 | 5.6 | 51.3 |
| τ³-Bench | 2.9 | 13.4 |
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.