| Benchmark | Claude 3.7 Sonnet | Nemotron 3 Super 120B A12B |
|---|---|---|
| AA Agentic Index | 37 | 8.8 |
| AA Intelligence | 27.6 | 25.7 |
| AA-LCR | 62.3 | 60.3 |
| AA-Omniscience | -0.7 | -41.5 |
| AIME 2025 | 54.8 | 90.2 |
| Artificial Analysis Coding Index | 36.4 | 37.7 |
| critpt | 0.9 | 3.1 |
| gdpval | 27.4 | 9.9 |
| GPQA Diamond | 84.8 | 82.7 |
| HLE | 9.7 | 22.8 |
| IFBench | 48.3 | 72.6 |
| multichallenge | 51.6 | 55.2 |
| OmniScience Accuracy | 28.2 | 24.3 |
| OmniScience Non-Hallucination | 60 | 13 |
| scicode | 40.3 | 42 |
| SWE-bench Verified | 70.3 | 53.7 |
| Terminal-Bench | 35.2 | 25.8 |
| Terminal-Bench Hard | 21.2 | 28.8 |
| τ²-Bench Telecom (AA run) | 54.7 | 67.8 |
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.