GPT-5.3 Codex (Medium): benchmark results
Provider: OpenAI. Released 2026-02-05. Access: API.
Aggregate Rating 1772 ± 26, rank #199 of 2170 rated models, from 22 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| CP-SynC-XL - MiniZinc + OR-Tools (Heuristic Prompt) | 54.1 | Correct instances (%; share of the 4,577 CP-SynC-XL instance | 100 |
| CP-SynC-XL - Python | 63.4 | Correct instances (%; share of the 4,577 CP-SynC-XL instance | 100 |
| MageBench S1 | 1717 | Combined 1v1 Elo | 91.2 |
| MageBench Season 1 | 1717 | Rating (imported) | 90.9 |
| CORE-Bench v1.1 (Codex CLI) | 97.4 | Task accuracy (%) on the 39 CORE-Bench v1.1 tasks (CORE-Benc | 85.7 |
| LLM Chess (Saplin) | 1053.1 | Native anchored-opponent Elo (signed points) | 83.1 |
| MageBench S2 | 1616 | Combined 1v1 Elo | 75 |
| ObviousBench | 97.22 | Answer pass³ (%) | 71.4 |
| DistDebug-Bench - Block | 59.7 | Coverage (%) | 66.7 |
| DistDebug-Bench - File | 79.1 | Coverage (%) | 66.7 |
| DistDebug-Bench - Block precision | 40 | Precision (%) | 58.3 |
| CP-SynC-XL - MiniZinc + OR-Tools | 52.2 | Correct instances (%; share of the 4,577 CP-SynC-XL instance | 50 |
Interactive version: theaggregate.ai/model?slug=gpt-5-3-codex-medium · How It Works · Data refreshed daily, snapshot 2026-11-05.