| Benchmark | Muse Spark | o3 |
|---|---|---|
| AA Agentic Index | 28.7 | 36.1 |
| AA Intelligence | 44.3 | 31.1 |
| AA-LCR | 77 | 73.3 |
| AA-Omniscience | 7.2 | -15.3 |
| ARC-AGI-2 | 42.5 | 6.5 |
| arena_vision | 1295 | 1217 |
| Artificial Analysis Coding Index | 58.6 | 38.4 |
| critpt | 11.3 | 1.1 |
| ERQA | 64.7 | 64 |
| Fortress | 20.2 | 16 |
| frontiermath_tier_4 | 14.6 | 2.1 |
| gdpval | 32.3 | 12.8 |
| GPQA Diamond | 89.5 | 83.3 |
| HLE | 58.4 | 20.3 |
| IFBench | 75.9 | 71.4 |
| MMMU-Pro | 80.5 | 76.4 |
| multichallenge | 75.5 | 60.4 |
| MultiNRC | 59 | 45.5 |
| OmniScience Accuracy | 49.6 | 38.6 |
| OmniScience Non-Hallucination | 26.8 | 12.9 |
| scicode | 51.5 | 41 |
| simpleqa_verified | 66.3 | 53 |
| SWE-bench Verified | 77.4 | 69.1 |
| Terminal-Bench Hard | 45.5 | 37.1 |
| τ²-Bench Telecom (AA run) | 91.5 | 80.7 |
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.