| Benchmark | o1 | Qwen3.6 27B |
|---|---|---|
| AA Agentic Index | 31.1 | 27.5 |
| AA Intelligence | 24 | 37.7 |
| AA-LCR | 63.3 | 73.3 |
| AA-Omniscience | -10.6 | -20 |
| Artificial Analysis Coding Index | 39.7 | 53.7 |
| critpt | 0.3 | 1.1 |
| gdpval | 11.5 | 32 |
| GPQA Diamond | 75.7 | 87.8 |
| HLE | 7.7 | 24 |
| humaneval | 88.1 | 78.2 |
| IFBench | 70.3 | 69.1 |
| MathVista | 71.8 | 87.4 |
| MMMU | 77.6 | 82.9 |
| OmniScience Accuracy | 34.7 | 19.6 |
| OmniScience Non-Hallucination | 30.7 | 50.7 |
| scicode | 35.8 | 39.8 |
| simpleqa | 47 | 95.7 |
| SWE-bench Verified | 48.9 | 77.2 |
| Terminal-Bench Hard | 12.9 | 34.8 |
| τ²-Bench Telecom (AA run) | 62.6 | 94.2 |
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.