| Benchmark | DeepSeek-V3 | Muse Spark 1.1 |
|---|---|---|
| AA Agentic Index | 1.6 | 39.7 |
| AA Intelligence | 15.4 | 53.2 |
| AA-LCR | 41.3 | 81.3 |
| AA-Omniscience | -37.6 | 28.1 |
| Artificial Analysis Coding Index | 23 | 71.3 |
| critpt | 0 | 15.1 |
| gdpval | 0 | 43.7 |
| GPQA Diamond | 68.4 | 89.8 |
| HLE | 5.2 | 62.1 |
| multichallenge | 31.4 | 75.3 |
| OmniScience Accuracy | 25.4 | 52 |
| OmniScience Non-Hallucination | 23.3 | 50 |
| scicode | 35.8 | 58.2 |
| TauBench V3 - Banking | 4.7 | 31.8 |
| Terminal-Bench 2.1 | 16.9 | 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.