| Benchmark | GPT-4o mini | GPT-5.6 Sol |
|---|---|---|
| AA Agentic Index | 1 | 57.8 |
| AA Intelligence | 7 | 61 |
| ARC-AGI-2 | 0 | 92.5 |
| Artificial Analysis Coding Index | 11.4 | 78.3 |
| gdpval | 0 | 61.1 |
| GPQA Diamond | 42.6 | 94.6 |
| HLE | 4.2 | 49.5 |
| IFBench | 31 | 72.7 |
| MMMU-Pro | 41.5 | 84.6 |
| scicode | 22.9 | 56.9 |
| simplebench | 10.7 | 64.8 |
| TauBench V3 - Banking | 2.9 | 44.3 |
| Terminal-Bench 2.1 | 5.6 | 89.5 |
| Video-MME | 68.9 | 89.5 |
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.