GPT-5.1-Codex-Max is a closed-weights model from OpenAI, released November 19, 2025.
| Domain | Benchmark | ||||
|---|---|---|---|---|---|
| Mathematics | ProofBenchBest score 100% | ||||
| Software engineering | METR Time HorizonsBest score 17 hr 25 min | 3 hr 44 min | 3 hr 44 min | 2 hr 2 min | 1 hr 50 min |
ALE-BenchBest score 2951 | 1209 | 1177 | 796 | — | |
Terminal-Bench 2.0Best score 85% |