| Benchmark | o3 | Qwen3.7 Plus Preview |
|---|---|---|
| AA Agentic Index | 36.1 | 20.8 |
| AA Intelligence | 31.1 | 39.4 |
| AA-LCR | 73.3 | 69 |
| AA-Omniscience | -15.3 | 2.4 |
| arena_vision | 1217 | 1263 |
| Artificial Analysis Coding Index | 38.4 | 55.9 |
| critpt | 1.1 | 9.1 |
| ERQA | 64 | 69.8 |
| gdpval | 12.8 | 22.2 |
| GPQA Diamond | 83.3 | 90.3 |
| HLE | 20.3 | 35.6 |
| IFBench | 71.4 | 79.1 |
| MMLU-Pro | 85 | 88.5 |
| MMMU-Pro | 76.4 | 80.5 |
| OmniScience Accuracy | 38.6 | 22.5 |
| OmniScience Non-Hallucination | 12.9 | 74.5 |
| scicode | 41 | 51.3 |
| SWE-bench Verified | 69.1 | 77.7 |
| Terminal-Bench Hard | 37.1 | 47 |
| VideoMMMU | 83.3 | 85.4 |
| τ²-Bench Telecom (AA run) | 80.7 | 93 |
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.