NVIDIA-Nemotron-Nano-9B-v2-Japanese (Thinking): benchmark results
Provider: NVIDIA. Released 2026-02-17. Access: Open.
Unified ELO 1570 ± 1, rank #837 of 3078 rated models, from 101 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Nejumi 4 - BFCL - Live AST | 74.07 | Accuracy (%) | 89.3 |
| Swallow - English MT-Bench - Math | 98.7 | Judge Score (normalized, %) | 81.3 |
| Nejumi 4 - Toxicity - Violation Categories | 48.41 | Criteria met (%) | 78.3 |
| Swallow - Japanese MT-Bench - Math | 97.8 | Judge Score (normalized, %) | 76.9 |
| Nejumi 4 - HalluLens | 96 | Hallucination resistance (%) | 75.4 |
| Nejumi 4 - BFCL - Irrelevance Detection | 86.67 | Accuracy (%) | 68.8 |
| Nejumi 4 - jaster (0-shot) - JSICK | 81 | Exact match (%) | 68 |
| Swallow - English MT-Bench - Writing | 73.7 | Judge Score (normalized, %) | 66.4 |
| Swallow - English MT-Bench - Reasoning | 83.5 | Judge Score (normalized, %) | 62.7 |
| Swallow - Post-trained Japanese - PolyMath High and Top | 39.6 | Accuracy (%) | 59.7 |
| Swallow - Post-trained English - IFBench | 43 | Instruction-Level Strict Accuracy (%) | 59 |
| Swallow - Japanese MT-Bench - Reasoning | 73.8 | Judge Score (normalized, %) | 56.7 |
Interactive version: theaggregate.ai/model?slug=nvidia-nemotron-nano-9b-v2-japanese-thinking · How It Works · Data refreshed daily, snapshot 2026-09-19.