Nex N2 Pro: benchmark results
Nex AGI's open 397B-A17B mixture-of-experts model for coding and agentic workflows. Provider: Nex AGI. Released 2026-06-03. Access: Open.
Aggregate Rating 1756 ± 15, rank #107 of 1656 rated models, from 138 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Nejumi 4 - MT-Bench (Japanese) - Writing | 99.5 | Judge rating (1-10, x10) | 98.6 |
| AA Omniscience - Software Engineering (SWE) - Rust | 79.17 | Accuracy (%) | 98.3 |
| Nejumi 4 - jaster (0-shot) - KUCI | 90 | Exact match (%) | 96.9 |
| Nejumi 4 - Toxicity - Social Norms | 99.21 | Criteria met (%) | 95.1 |
| AA Omniscience - Software Engineering (SWE) - JavaScript | 53.7 | Accuracy (%) | 94.9 |
| Nejumi 4 - jaster (2-shot) - JCommonsenseQA | 100 | Exact match (%) | 94.8 |
| Nejumi 4 - jaster (2-shot) - JSICK | 85 | Exact match (%) | 94.8 |
| AA Omniscience - Software Engineering (SWE) - Go | 42.86 | Accuracy (%) | 94.5 |
| AA Omniscience - Software Engineering (SWE) - C | 67.35 | Accuracy (%) | 94.2 |
| Nejumi 4 - jaster (2-shot) - JSQuAD | 81.56 | Character F1 (x100) | 94.1 |
| AA Omniscience - Software Engineering (SWE) - TypeScript | 48.89 | Accuracy (%) | 93.3 |
| Nejumi 4 - jaster (2-shot) - JBLiMP | 96 | Exact match (%) | 91.6 |
Interactive version: theaggregate.ai/model?slug=nex-n2-pro · How It Works · Data refreshed daily, snapshot 2026-10-20.