GPT-5.2 Codex (xHigh): benchmark results
GPT-5.2 Codex evaluated at the xhigh reasoning-effort setting. Provider: OpenAI. Released 2025-12-18. Access: API.
Aggregate Rating 1759 ± 23, rank #215 of 2168 rated models, from 48 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Context-Bench Filesystem | 93 | Rubric Score (%) | 100 |
| ResearchGym - Cross-Modal Retrieval | 96.6 | Best-of-3 score normalised to the task's reference (SOTA) re | 100 |
| ResearchGym - Improving Replay Buffers | 49.7 | Best-of-3 score normalised to the task's reference (SOTA) re | 100 |
| AA IFBench | 77.62 | Accuracy (%) | 96.4 |
| BinaryAudit | 90.67 | Avg Success Rate (%) | 96 |
| AA Long Context Reasoning | 82.33 | Accuracy (%) | 92.8 |
| ALE-Bench | 1299.9 | Performance (Self-Refine x1) (imported) | 91.1 |
| AA GPQA Diamond | 89.9 | Accuracy (%) | 89.2 |
| AA TAU-2 Bench | 92.11 | Accuracy (%) | 87.5 |
| Tau2-Bench Telecom | 92.11 | Success Rate (%) | 87.2 |
| AA Terminal-Bench Hard | 37.12 | Accuracy (%) | 83.9 |
| AA Humanity's Last Exam | 35.73 | Accuracy (%) | 82.8 |
Interactive version: theaggregate.ai/model?slug=gpt-5-2-codex-xhigh · How It Works · Data refreshed daily, snapshot 2026-10-20.