Across 15 shared benchmarks, Nemotron 3 Super 120B A12B scores higher on 1 and nvidia-nemotron-3-ultra-550b-a55b on 14. The widest gap is gdpval, where nvidia-nemotron-3-ultra-550b-a55b scores 46.7 against 9.9. Nemotron 3 Super 120B A12B is the cheaper of the two on tracked API pricing ($0.20 against $0.60 per million input tokens).
| Benchmark | Nemotron 3 Super 120B A12B | nvidia-nemotron-3-ultra-550b-a55b |
|---|---|---|
| Airline | 56.9 | 81.5 |
| browsecomp | 31.3 | 44.4 |
| gdpval | 9.9 | 46.7 |
| GPQA Diamond | 82.7 | 87 |
| HLE | 22.8 | 26.7 |
| MMLU-Pro | 83.8 | 86.8 |
| multichallenge | 55.2 | 63.8 |
| Retail | 64.3 | 86.4 |
| ruler | 91.8 | 94.7 |
| SciCode (subtask) | 42.3 | 44.6 |
| SWE-bench Multilingual | 45.8 | 67.7 |
| SWE-bench Verified | 53.7 | 70.7 |
| Telecom | 61.1 | 92.9 |
| Terminal-Bench 2.1 | 38.6 | 56.4 |
| WMT24++ (en→xx) | 86.8 | 83.7 |
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. Quoted rates are the price-setter row we currently track for each model — its direct or vendor-official listing where one exists (direct, direct), otherwise the lowest tracked offer.