| Benchmark | Claude 3.5 Sonnet | Qwen3.6 Plus |
|---|---|---|
| AA Intelligence | 10 | 40.5 |
| AI2D | 94.7 | 94.4 |
| Artificial Analysis Coding Index | 30.2 | 54.5 |
| C-Eval | 76.7 | 93.3 |
| frontiermath_tier_4 | 0 | 8.3 |
| GPQA Diamond | 67.2 | 90.4 |
| HLE | 3.9 | 28.8 |
| ifeval | 90.1 | 94.3 |
| LiveCodeBench v6 | 37.2 | 87.1 |
| longbench_v2 | 41 | 62 |
| mmlu_redux | 88.9 | 94.5 |
| MMLU-Pro | 78 | 88.5 |
| MMMU | 72 | 86 |
| MMMU-Pro | 54.7 | 78.8 |
| MMStar | 62.2 | 83.3 |
| RealWorldQA | 60.1 | 85.4 |
| scicode | 36.6 | 40.7 |
| supergpqa | 48.2 | 71.6 |
| SWE-bench Verified | 50.8 | 78.8 |
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.