GPT-5.6 Luna (Non-reasoning): benchmark results
GPT-5.6 Luna evaluated with reasoning disabled. Provider: OpenAI. Released 2026-07-09. Access: API.
Unified ELO 1587 ± 1, rank #485 of 1761 rated models, from 25 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| UGI - Natural Intelligence | 45.85 | NatInt Score | 88.6 |
| AA Omniscience - Software Engineering (SWE) | 49.9 | Accuracy (%) | 80.3 |
| UGI - Writing | 40.98 | Writing Score | 76.3 |
| AA-Omniscience Accuracy | 28.6 | Accuracy (%) | 71.5 |
| AA Omniscience - Law | 18.4 | Accuracy (%) | 68.7 |
| AA Omniscience - Business | 21.8 | Accuracy (%) | 68 |
| AA Omniscience - Health | 25 | Accuracy (%) | 66.6 |
| Artificial Analysis Intelligence Index | 19.32 | Intelligence Index | 65.1 |
| AA Omniscience - Science, Engineering & Mathematics | 32 | Accuracy (%) | 63.6 |
| AA Omniscience - Humanities & Social Sciences | 24.5 | Accuracy (%) | 61.8 |
| FrontierMath - Tiers 1-3 (v2) | 39.65 | Accuracy (%, 285 private v2 problems) | 60.6 |
| AA Omniscience | -24.95 | Score | 60.4 |
Interactive version: theaggregate.ai/model?slug=gpt-5-6-luna-non-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-05.