| Benchmark | Claude 3.5 Sonnet | Muse Spark |
|---|---|---|
| AA Intelligence | 10 | 44.3 |
| arena_vision | 1146 | 1295 |
| Artificial Analysis Coding Index | 30.2 | 58.6 |
| Fortress | 13 | 20.2 |
| frontiermath_tier_4 | 0 | 14.6 |
| GPQA Diamond | 67.2 | 89.5 |
| HLE | 3.9 | 58.4 |
| MMMU-Pro | 54.7 | 80.5 |
| scicode | 36.6 | 51.5 |
| SWE-bench Verified | 50.8 | 77.4 |
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.