| Benchmark | Claude 3.5 Sonnet | Qwen3.7 Plus Preview |
|---|---|---|
| AA Intelligence | 10 | 39.4 |
| arena_vision | 1146 | 1263 |
| Artificial Analysis Coding Index | 30.2 | 55.9 |
| GPQA Diamond | 67.2 | 90.3 |
| HLE | 3.9 | 35.6 |
| ifeval | 90.1 | 94.6 |
| LiveCodeBench v6 | 37.2 | 89.6 |
| mmlu_redux | 88.9 | 94.5 |
| MMLU-Pro | 78 | 88.5 |
| MMMU-Pro | 54.7 | 80.5 |
| RealWorldQA | 60.1 | 86.9 |
| scicode | 36.6 | 51.3 |
| supergpqa | 48.2 | 71.4 |
| SWE-bench Verified | 50.8 | 77.7 |
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.