NVIDIA-Nemotron-3-Nano-30B-A3B-BF16: benchmark results
NVIDIA's Nemotron-3-Nano-30B-A3B-BF16, a hybrid Mamba-2/MoE model (3.5B active of 30B total, 128 experts) with configurable reasoning traces, supporting English, German, Spanish, French, Italian and Japanese. Provider: NVIDIA. Released 2025-12-15. Access: Open.
Aggregate Rating 1531 ± 13, rank #673 of 1635 rated models, from 523 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EuroEval Dutch NLU - CoNLL NL | 80.79 | Micro F1 excluding MISC (%) | 99.7 |
| EuroEval English NLU - SQuAD | 92.09 | Character F1 (%) | 99.7 |
| EuroEval Polish Summarization - PSC | 25.68 | chrF3++ (%) | 98.6 |
| EuroEval French NLU - FQuAD | 76.26 | Character F1 (%) | 98.1 |
| EuroEval German NLU - GermEval | 75.4 | Micro F1 excluding MISC (%) | 98.1 |
| EuroEval Slovene NLU - SSJ500K NER | 76.38 | Micro F1 excluding MISC (%) | 97.8 |
| EuroEval Portuguese NLU | 66.72 | NLU Average Score (%) | 97.5 |
| EuroEval English NLU - ScaLA EN | 67.64 | Matthews correlation coefficient (%) | 97 |
| EuroEval Italian NLU | 67.36 | NLU Average Score (%) | 96.5 |
| EuroEval French NLU - Eltec | 73.31 | Micro F1 excluding MISC (%) | 95.7 |
| EuroEval Italian NLU - ScaLA IT | 57.37 | Matthews correlation coefficient (%) | 95.6 |
| EuroEval English NLU | 74.33 | NLU Average Score (%) | 95.2 |
Interactive version: theaggregate.ai/model?slug=nvidia-nemotron-3-nano-30b-a3b-bf16 · How It Works · Data refreshed daily, snapshot 2026-10-15.