| Benchmark | Muse Spark | o1 |
|---|---|---|
| AA Agentic Index | 28.7 | 31.1 |
| AA Intelligence | 44.3 | 24 |
| AA-LCR | 77 | 63.3 |
| AA-Omniscience | 7.2 | -10.6 |
| arena_vision | 1295 | 1193 |
| Artificial Analysis Coding Index | 58.6 | 39.7 |
| critpt | 11.3 | 0.3 |
| Fortress | 20.2 | 19.4 |
| gdpval | 32.3 | 11.5 |
| GPQA Diamond | 89.5 | 75.7 |
| HLE | 58.4 | 7.7 |
| IFBench | 75.9 | 70.3 |
| OmniScience Accuracy | 49.6 | 34.7 |
| OmniScience Non-Hallucination | 26.8 | 30.7 |
| scicode | 51.5 | 35.8 |
| SWE-bench Verified | 77.4 | 48.9 |
| Terminal-Bench Hard | 45.5 | 12.9 |
| τ²-Bench Telecom (AA run) | 91.5 | 62.6 |
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.