← All benchmarks
Nemotron 3 Super 120B A12B (Reasoning)
NVIDIA · Released 2026-03-11 · Independent benchmark scores from Artificial Analysis
Intelligence Index
18.6
Output speed
141.9
tokens/sec
Blended price
$0.35
per 1M tokens
Time to first token
1.12s
Model metadata
| Model ID | 63872e9c-3377-4a6b-b477-7bba244c38e9 |
|---|---|
| Slug | nvidia-nemotron-3-super-120b-a12b |
| Release date | 2026-03-11 |
| Provider | NVIDIA |
| Provider slug | nvidia |
| Output speed | 141.9 tok/s |
| Time to first token | 1.12s |
| Time to first answer | 15.21s |
| Blended price ($/1M) | $0.35 |
| Input price ($/1M) | $0.20 |
| Output price ($/1M) | $0.80 |
All evaluation scores (17)
| aime | — |
|---|---|
| aime 25 | — |
| artificial analysis coding index | 37.70 |
| artificial analysis intelligence index | 18.60 |
| artificial analysis math index | — |
| gpqa | 80.0% |
| hle | 20.8% |
| ifbench | 71.5% |
| lcr | 65.7% |
| livecodebench | — |
| math 500 | — |
| mmlu pro | — |
| scicode | — |
| tau banking | 10.3% |
| tau2 | 67.8% |
| terminalbench hard | 28.8% |
| terminalbench v2 1 | 38.6% |
Source: artificialanalysis.ai/models/nvidia-nemotron-3-super-120b-a12b
Benchmark data from Artificial Analysis.