| Benchmark | Nemotron 3 Super 120B A12B | Qwen3.7 Max |
|---|---|---|
| AA Agentic Index | 8.8 | 30.9 |
| AA Intelligence | 25.7 | 46.7 |
| AA-LCR | 60.3 | 74.7 |
| AA-Omniscience | -41.5 | 13.5 |
| aime_2026 | 91.7 | 97 |
| Artificial Analysis Coding Index | 37.7 | 66 |
| critpt | 3.1 | 13.4 |
| gdpval | 9.9 | 38.5 |
| GPQA Diamond | 82.7 | 92.4 |
| HLE | 22.8 | 41.4 |
| hmmt_feb_2026 | 85.2 | 97.1 |
| IFBench | 72.6 | 80.5 |
| itbenchSre | 1.1 | 42.5 |
| mmlu_prox | 79.4 | 87 |
| MMLU-Pro | 83.8 | 89.6 |
| OmniScience Accuracy | 24.3 | 31.1 |
| OmniScience Non-Hallucination | 13 | 74.4 |
| scicode | 42 | 53.5 |
| SWE-bench Multilingual | 45.8 | 78.3 |
| SWE-bench Verified | 53.7 | 80.4 |
| TauBench V3 - Banking | 10.3 | 11.8 |
| Terminal-Bench 2.0 | 31 | 69.7 |
| Terminal-Bench 2.1 | 38.6 | 75 |
| Terminal-Bench Hard | 28.8 | 50.8 |
| τ²-Bench Telecom (AA run) | 67.8 | 94.7 |
| τ³-Bench | 10.1 | 10.9 |
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.