| Benchmark | GPT-5.1 | o1 |
|---|---|---|
| AA Agentic Index | 32.2 | 31.1 |
| AA Intelligence | 37.5 | 24 |
| AA-LCR | 76.7 | 63.3 |
| AA-Omniscience | 5.4 | -10.6 |
| AIME 2025 | 94.2 | 81.7 |
| arena_vision | 1250 | 1193 |
| Artificial Analysis Coding Index | 49.4 | 39.7 |
| critpt | 4.9 | 0.3 |
| Fortress | 25.7 | 19.4 |
| gdpval | 25 | 11.5 |
| GPQA Diamond | 88.1 | 75.7 |
| HLE | 28.5 | 7.7 |
| IFBench | 72.9 | 70.3 |
| mmmlu | 91 | 87.7 |
| MMMU | 85.4 | 77.6 |
| OmniScience Accuracy | 37.7 | 34.7 |
| OmniScience Non-Hallucination | 48.1 | 30.7 |
| scicode | 43.3 | 35.8 |
| SEAL VISTA | 43.8 | 45.3 |
| simplebench | 53.2 | 41.7 |
| SWE-bench Verified | 76.3 | 48.9 |
| Terminal-Bench Hard | 45.5 | 12.9 |
| τ²-Bench Telecom (AA run) | 81.9 | 62.6 |
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.