GPT-5 Nano (2025-08-07) (High): benchmark results
Provider: OpenAI. Released 2025-08-07. Access: API.
Unified ELO 1587 ± 14, rank #728 of 2131 rated models, from 202 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Nejumi 4 - jaster (0-shot) - JSICK | 87 | Exact match (%) | 99.6 |
| Nejumi 4 - jaster (2-shot) - JSICK | 88 | Exact match (%) | 99.3 |
| Nejumi 4 - Toxicity - Social Norms | 99.6 | Criteria met (%) | 98.6 |
| EuroEval English Knowledge | 94.92 | Knowledge Average Score (%) | 96.8 |
| EuroEval French NLU - Allocine | 96.47 | Sentiment classification Score (%) | 95.3 |
| EuroEval Dutch Knowledge | 82.77 | Knowledge Average Score (%) | 94.7 |
| EuroEval Norwegian Common Sense Reasoning | 86.99 | Common Sense Reasoning Average Score (%) | 94.4 |
| EuroEval Finnish | 65.28 | Average Score (%) | 94.1 |
| EuroEval Icelandic NLU - ScaLA IS | 44.69 | Linguistic acceptability Score (%) | 93.7 |
| EuroEval Swedish Knowledge | 79.04 | Knowledge Average Score (%) | 92.9 |
| EuroEval French Knowledge | 79.05 | Knowledge Average Score (%) | 92.7 |
| EuroEval Norwegian Knowledge - Idioms NO | 52.59 | MCC (x100) | 92.2 |
Interactive version: theaggregate.ai/model?slug=gpt-5-nano-2025-08-07-high · How It Works · Data refreshed daily, snapshot 2026-10-09.