Nemotron 3.5 Super 120B-A12B Base places in the top three on RACE.
NVIDIA Oct 2026 Open weights Flagship
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| ARC-Challenge | 96% | #5 of 49 | Lab model card | |
| TruthfulQA | 71.2% | #5 of 30 | Lab model card | |
| PIQA | 85.1% | #7 of 37 | Lab model card | |
| HellaSwag | 89% | #11 of 53 | Lab model card | |
| CommonSenseQA | 85.3% | #4 of 19 | Lab model card | |
| AGIEval | 75.7% | #10 of 40 | Lab model card | |
| MGSM | 87.5% | #18 of 57 | Lab model card | |
| RACE | 91.4% | #3 of 8 | Lab model card | |
| MMLU | 84.1% | #58 of 131 | Lab model card | |
| WinoGrande | 79.1% | #23 of 47 | Lab model card | |
| MMLU-Pro | 76.3% | #104 of 167 | Lab model card | |
| OpenBookQA | 48.8% | #17 of 24 | Lab model card | |
| Global MMLU | 86.6% | #11 of 14 | Lab model card | |
| SIQA | 49% | #21 of 24 | Lab model card | |
| SuperGPQA | 47.6% | #31 of 35 | Lab model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MATH | 87.7% | #13 of 99 | Lab model card | |
| GSM8K | 91.7% | #34 of 92 | Lab model card | |
| GPQA Diamond | 66.2% | #199 of 302 | Lab model card | |
| MATH-500 | 81.6% | #61 of 70 | Lab model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MBPP | 79.1% | #27 of 71 | Lab model card | |
| BigCodeBench | 48.4% | #7 of 14 | Lab model card | |
| HumanEval | 81.3% | #52 of 101 | Lab model card | |
| LiveCodeBench | 47.6% | #32 of 54 | Lab model card |
The Self-Reported Index fits one score to the 23 capability results NVIDIA reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 124.9 to 131.3. How we compute it.
Every result in the longer documents the lab published with it. 23 of them aren't in the launch post.