| Benchmark | GPT-4o mini | Nemotron 3 Super 120B A12B |
|---|---|---|
| AA Agentic Index | 1 | 8.8 |
| AA Intelligence | 7 | 25.7 |
| Artificial Analysis Coding Index | 11.4 | 37.7 |
| gdpval | 0 | 9.9 |
| GPQA Diamond | 42.6 | 82.7 |
| GSM8K | 84.3 | 90.7 |
| HLE | 4.2 | 22.8 |
| humaneval | 87.2 | 80.5 |
| IFBench | 31 | 72.6 |
| LiveCodeBench | 35.5 | 82.1 |
| mmlu | 81.8 | 86 |
| OpenBookQA | 92 | 48.6 |
| scicode | 22.9 | 42 |
| SWE-bench Verified | 8.7 | 53.7 |
| TauBench V3 - Banking | 2.9 | 10.3 |
| Terminal-Bench 2.1 | 5.6 | 38.6 |
| τ³-Bench | 2.9 | 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.