GPT-5
Benchmark scores
| Benchmark | Category | Rank | Score | Captured |
|---|---|---|---|---|
| Aider Polyglot | code | #1 | 88.0% | 2025-08-23 |
| AA Coding Index | code | #1 | 78.3% | 2026-07-31 |
| Terminal-Bench 2.0 | agents | #4 | 88.1% | 2026-04-30 |
| Chatbot Arena | chat | #6 | 1469 | 2026-07-27 |
| OpenRouter · Weekly Usage | usage | #8 | #20 | 2026-06-09 |
| SWE-bench Verified | agents | #9 | 74.4% | 2025-10-15 |
| METR Time Horizon | agents | #9 | 3h23m | 2025-08-07 |