| Benchmark | o1 | Qwen3.6 35B A3B |
|---|---|---|
| AA Agentic Index | 31.1 | 21.6 |
| AA Intelligence | 24 | 32.1 |
| AA-LCR | 63.3 | 66.7 |
| AA-Omniscience | -10.6 | -22.2 |
| Artificial Analysis Coding Index | 39.7 | 41.9 |
| critpt | 0.3 | 0.3 |
| gdpval | 11.5 | 27.8 |
| GPQA Diamond | 75.7 | 86 |
| HLE | 7.7 | 22.2 |
| IFBench | 70.3 | 64.4 |
| MathVista | 71.8 | 86.4 |
| MMMU | 77.6 | 81.7 |
| OmniScience Accuracy | 34.7 | 18.8 |
| OmniScience Non-Hallucination | 30.7 | 49.5 |
| scicode | 35.8 | 35.8 |
| SWE-bench Verified | 48.9 | 73.4 |
| Terminal-Bench Hard | 12.9 | 34.8 |
| τ²-Bench Telecom (AA run) | 62.6 | 95.3 |
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.