GPT-5.2 Codex (High): benchmark results

GPT-5.2 Codex evaluated at the high reasoning-effort setting. Provider: OpenAI. Released 2025-12-18. Access: API.

Aggregate Rating 1737 ± 24, rank #299 of 2284 rated models, from 31 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
SWE-AGI (Hard) - Test-Suite Pass Rate91.2Test-suite pass rate (%, 0-100): mean over the 8 hard-tier S100
BinaryAudit88.41Avg Success Rate (%)88
SWE-AGI (Easy)6Tasks passed (out of 6): the easy-tier SWE-AGI tasks whose f83.3
SWE-AGI (Easy) - Test-Suite Pass Rate100Test-suite pass rate (%, 0-100): mean over the 6 easy-tier S83.3
SlopCodeBench (Just Solve) - Isolated21.9Isolated solve rate (%): share of the 196 checkpoints whose 82.4
APEX-Agents42.2Mean Score (ReAct) (imported)79.5
WeaveBench - Data Analysis and Visualization15.4PassRate (%) on the Data Analysis and Visualization domain t72.2
SlopCodeBench (Just Solve)9.7Strict solve rate (%): share of the 196 checkpoints whose wo70.6
SWE-AGI (Medium)7Tasks passed (out of 8): the medium-tier SWE-AGI tasks whose66.7
SWE-AGI (Medium) - Test-Suite Pass Rate98.9Test-suite pass rate (%, 0-100): mean over the 8 medium-tier66.7
WeaveBench - DevOps16.7PassRate (%) on the DevOps domain tasks of 114 long-horizon 61.1
Sonar LLM Leaderboard - Java - Pass Rate80.33Passing tests (%)60.2

Interactive version: theaggregate.ai/model?slug=gpt-5-2-codex-high · How It Works · Data refreshed daily, snapshot 2026-11-21.