| Benchmark | DeepSeek-R1 | Muse Spark 1.1 |
|---|---|---|
| AA Agentic Index | 3.1 | 39.7 |
| AA Intelligence | 18.6 | 53.2 |
| AA-LCR | 56 | 81.3 |
| AA-Omniscience | -31.3 | 28.1 |
| Artificial Analysis Coding Index | 24.6 | 71.3 |
| critpt | 0.6 | 15.1 |
| Fortress | 74.4 | 12.4 |
| gdpval | 1.5 | 43.7 |
| GPQA Diamond | 81 | 89.8 |
| HLE | 17.7 | 62.1 |
| multichallenge | 45 | 75.3 |
| MultiNRC | 27.6 | 65.6 |
| OmniScience Accuracy | 30.7 | 52 |
| OmniScience Non-Hallucination | 10.5 | 50 |
| scicode | 35.7 | 58.2 |
| TauBench V3 - Banking | 6.4 | 31.8 |
| Terminal-Bench 2.1 | 19.1 | 80 |
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.