| Benchmark | Muse Spark | Nemotron 3 Super 120B A12B |
|---|---|---|
| AA Agentic Index | 28.7 | 8.8 |
| AA Intelligence | 44.3 | 25.7 |
| AA-LCR | 77 | 60.3 |
| AA-Omniscience | 7.2 | -41.5 |
| Artificial Analysis Coding Index | 58.6 | 37.7 |
| critpt | 11.3 | 3.1 |
| gdpval | 32.3 | 9.9 |
| GPQA Diamond | 89.5 | 82.7 |
| HLE | 58.4 | 22.8 |
| IFBench | 75.9 | 72.6 |
| itbenchSre | 16.2 | 1.1 |
| multichallenge | 75.5 | 55.2 |
| OmniScience Accuracy | 49.6 | 24.3 |
| OmniScience Non-Hallucination | 26.8 | 13 |
| scicode | 51.5 | 42 |
| SWE-bench Verified | 77.4 | 53.7 |
| TauBench V3 - Banking | 19.6 | 10.3 |
| Terminal-Bench 2.0 | 59 | 31 |
| Terminal-Bench 2.1 | 62.2 | 38.6 |
| Terminal-Bench Hard | 45.5 | 28.8 |
| τ²-Bench Telecom (AA run) | 91.5 | 67.8 |
| τ³-Bench | 19.6 | 10.1 |
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.