Nemotron 3 Nano 30B A3B (Reasoning): benchmark results
Provider: NVIDIA. Released 2025-12-15. Access: Open.
Unified ELO 1510 ± 13, rank #989 of 2088 rated models, from 95 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA AIME 2025 | 91 | Accuracy (%) | 92.1 |
| AA IFBench | 71.09 | Accuracy (%) | 86.4 |
| AA LiveCodeBench | 74.07 | Pass@1 (%) | 85.7 |
| CoCoReviewBench - Category Completeness | 82.93 | Category completeness (%): the average number of review cate | 82.4 |
| CoCoReviewBench - Clarity | 0.36 | Category-level clarity score (coherence, readability and str | 76.5 |
| CoCoReviewBench - Correctness | -0.01 | Category-level correctness score (agreement of each review c | 67.6 |
| AA MMLU-Pro | 79.38 | Accuracy (%) | 64.1 |
| AA Omniscience - Software Engineering (SWE) - Kotlin | 20 | Accuracy (%) | 63 |
| AA Omniscience - Software Engineering (SWE) - Swift | 36 | Accuracy (%) | 60.6 |
| NLCO - Set-L - Accuracy | 38.2 | Accuracy: feasible and matching the solver-optimal objective | 60 |
| NLCO - Set-M - Feasibility Rate | 88.2 | Average feasibility rate: output satisfies every constraint | 60 |
| CoCoReviewBench | 0.07 | Average category-level review score (the mean of correctness | 58.8 |
Interactive version: theaggregate.ai/model?slug=nemotron-3-nano-30b-a3b-reasoning · How It Works · Data refreshed daily, snapshot 2026-10-07.