| Benchmark | Claude 3.5 Sonnet | GLM-5.1 |
|---|---|---|
| AA Intelligence | 10 | 41 |
| Artificial Analysis Coding Index | 30.2 | 55.8 |
| frontiermath_tier_4 | 0 | 12.5 |
| GPQA Diamond | 67.2 | 86.8 |
| HLE | 3.9 | 52.3 |
| MMLU-Pro | 78 | 86 |
| scicode | 36.6 | 43.8 |
| simplebench | 41.4 | 55.1 |
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.