| Benchmark | o3 | Qwen3.7 Max |
|---|---|---|
| AA Agentic Index | 36.1 | 30.9 |
| AA Intelligence | 31.1 | 46.7 |
| AA-LCR | 73.3 | 74.7 |
| AA-Omniscience | -15.3 | 13.5 |
| Artificial Analysis Coding Index | 38.4 | 66 |
| critpt | 1.1 | 13.4 |
| gdpval | 12.8 | 38.5 |
| GPQA Diamond | 83.3 | 92.4 |
| HLE | 20.3 | 41.4 |
| IFBench | 71.4 | 80.5 |
| longbench_v2 | 58.8 | 65.3 |
| MMLU-Pro | 85 | 89.6 |
| OmniScience Accuracy | 38.6 | 31.1 |
| OmniScience Non-Hallucination | 12.9 | 74.4 |
| scicode | 41 | 53.5 |
| simplebench | 53.1 | 70.4 |
| simpleqa_verified | 53 | 58.5 |
| SWE-bench Verified | 69.1 | 80.4 |
| Terminal-Bench Hard | 37.1 | 50.8 |
| τ²-Bench Telecom (AA run) | 80.7 | 94.7 |
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.