GPT-5.2 Codex: benchmark results
OpenAI's coding-specialized GPT-5.2 Codex model for software tasks. Provider: OpenAI. Released 2026-01-14. Access: API.
Unified ELO 1689 ± 1, rank #71 of 1392 rated models, from 60 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| LiveBench Code Completion | 86.96 | Score | 99.1 |
| LM Market Cap LMC Score | 90.5 | LMC Score (0-100) | 94.5 |
| Vals AI LiveCodeBench | 87.99 | Accuracy (%) | 94.4 |
| SnakeBench | 32.4 | TrueSkill Rating | 91.6 |
| BenchmarkList ECI | 140.73 | Capability Index (ECI) | 89.9 |
| AISI Cyber CTF | 72 | Success Rate (%) | 89.3 |
| AI Chess Leaderboard (Reasoning) | 1303 | Elo | 88.2 |
| AI Chess Leaderboard (Continuation) | 1151 | Elo | 86 |
| Terminal-Bench 2.0 | 66.5 | Accuracy (%) | 85.7 |
| SWE-bench Verified | 72.8 | Resolved (%) | 84.8 |
| LiveBench Table Reformat | 100 | Score | 83 |
| LiveBench Consecutive Events | 89.02 | Score | 77.4 |
Interactive version: theaggregate.ai/model?slug=gpt-5-2-codex · How It Works · Data refreshed daily, snapshot 2026-09-05.