| Benchmark | Claude Sonnet 4 | Muse Spark 1.1 |
|---|---|---|
| AA Agentic Index | 39.2 | 39.7 |
| AA Intelligence | 29.8 | 53.2 |
| AA-LCR | 69.7 | 81.3 |
| AA-Omniscience | 0.2 | 28.1 |
| Artificial Analysis Coding Index | 37.6 | 71.3 |
| critpt | 1.1 | 15.1 |
| Fortress | 24.4 | 12.4 |
| gdpval | 31.2 | 43.7 |
| GPQA Diamond | 78 | 89.8 |
| HLE | 10.7 | 62.1 |
| multichallenge | 57.1 | 75.3 |
| MultiNRC | 18.4 | 65.6 |
| OmniScience Accuracy | 22.7 | 52 |
| OmniScience Non-Hallucination | 70.9 | 50 |
| OSWorld-Verified | 42.2 | 80.8 |
| scicode | 40 | 58.2 |
| SWE-bench Pro | 42.7 | 61.5 |
| TauBench V3 - Banking | 16.7 | 31.8 |
| Terminal-Bench 2.1 | 36.3 | 80 |
| VisualToolBench | 4.5 | 44.8 |
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.