| Benchmark | Muse Spark 1.1 | o3 |
|---|---|---|
| AA Agentic Index | 39.7 | 36.1 |
| AA Intelligence | 53.2 | 31.1 |
| AA-LCR | 81.3 | 73.3 |
| AA-Omniscience | 28.1 | -15.3 |
| Artificial Analysis Coding Index | 71.3 | 38.4 |
| critpt | 15.1 | 1.1 |
| Fortress | 12.4 | 16 |
| gdpval | 43.7 | 12.8 |
| GPQA Diamond | 89.8 | 83.3 |
| HLE | 62.1 | 20.3 |
| multichallenge | 75.3 | 60.4 |
| MultiNRC | 65.6 | 45.5 |
| OmniScience Accuracy | 52 | 38.6 |
| OmniScience Non-Hallucination | 50 | 12.9 |
| scicode | 58.2 | 41 |
| VisualToolBench | 44.8 | 13.7 |
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.