Claude Opus 4 (Non-reasoning): benchmark results
Claude Opus 4 evaluated with reasoning disabled. Provider: Anthropic. Released 2025-05-22. Access: API.
Aggregate Rating 1667 ± 21, rank #468 of 2168 rated models, from 86 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Nejumi 4 - jaster (2-shot) - AIO | 96.86 | Character F1 (x100) | 100 |
| Nejumi 4 - jaster (0-shot) - JNLI | 93 | Exact match (%) | 99.7 |
| Nejumi 4 - jaster (2-shot) - JNLI | 94 | Exact match (%) | 99 |
| Nejumi 4 - jaster (0-shot) - AIO | 95.19 | Character F1 (x100) | 97.2 |
| Nejumi 4 - jaster (0-shot) - KUCI | 90 | Exact match (%) | 96.9 |
| UGI - Writing | 68.24 | Writing Score | 96.9 |
| Nejumi 4 - BFCL - Non-Live AST | 82.59 | Accuracy (%) | 96.5 |
| Nejumi 4 - HalluLens | 100 | Hallucination resistance (%) | 95.8 |
| Nejumi 4 - MT-Bench (Japanese) - Extraction | 98.5 | Judge rating (1-10, x10) | 95.8 |
| Nejumi 4 - jaster (0-shot) - ALT English-to-Japanese | 92.54 | COMET wmt22 (x100) | 95.8 |
| Generalization V1 (Lechmazur) | 1.7 | Avg Rank (lower is better) | 95.6 |
| Nejumi 4 - JBBQ (2-shot) - Accuracy | 95 | Accuracy (%) | 95.5 |
Interactive version: theaggregate.ai/model?slug=claude-opus-4-non-reasoning · How It Works · Data refreshed daily, snapshot 2026-10-20.