GPT-5.6 (Medium) — benchmark results
Provider: OpenAI. Access: API.
Unified ELO 2012 ± 37, rank #21 of 1723 rated models, from 38 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA Omniscience - Business | 50.9 | Accuracy (%) | 99.8 |
| AA Terminal-Bench Hard | 62.88 | Accuracy (%) | 99.5 |
| AA Omniscience - Software Engineering (SWE) - Go | 84 | Accuracy (%) | 99.2 |
| AA Omniscience - Software Engineering (SWE) - Rust | 86 | Accuracy (%) | 99.2 |
| AA SciCode | 56.48 | Accuracy (%) | 99.2 |
| AA Omniscience - Humanities & Social Sciences | 56.9 | Accuracy (%) | 99.1 |
| AA Omniscience - Software Engineering (SWE) - R | 72 | Accuracy (%) | 98.8 |
| AA-Omniscience Accuracy | 56.68 | Accuracy (%) | 98.8 |
| AA Omniscience - Software Engineering (SWE) - TypeScript | 90 | Accuracy (%) | 98.7 |
| Artificial Analysis Intelligence Index | 53.59 | Intelligence Index | 98.7 |
| AA Omniscience - Science, Engineering & Mathematics | 51.9 | Accuracy (%) | 98.5 |
| AA Omniscience - Health | 48.6 | Accuracy (%) | 98.4 |
Interactive version: aibenchmarks.dev/model?slug=gpt-5-6-medium · How the rankings work · Data refreshed daily, snapshot 2026-07-10.