evals.report
BenchmarksLabsCompareRun guidesIn the wild

GPT-5.2-Codex

OpenAI · GPT. Released Dec 18, 2025.

GPT-5.2-Codex is a model from OpenAI in the GPT family, released Dec 18, 2025. evals.report tracks 8 reported GPT-5.2-Codex benchmark scores across SWE-bench Pro, PostTrainBench, Artificial Analysis Intelligence Index, GDPval, SciCode, WebDev Arena, SWE-bench Multilingual, Vibe Code Bench — each shown with its benchmark, metric, source status, and date, and never combined into a single ranking.

8 results

Benchmark results 8

Compare this model
BenchmarkCategoryScoreMetricStatusDate
SWE-bench ProCoding41.04%% resolvedOfficialDec 18, 2025Details
PostTrainBenchAgents17.22%weighted average scoreOfficialDec 18, 2025Details
Artificial Analysis Intelligence IndexReasoning49IndexUnverifiedDec 18, 2025Details
GDPvalAgents1288EloOfficialDec 18, 2025Details
SciCodeCoding54.6%accuracyUnverifiedDec 18, 2025Details
WebDev ArenaChat preference1335EloVerifiedDec 18, 2025Details
SWE-bench MultilingualCoding66.3%% resolvedOfficialDec 18, 2025Details
Vibe Code BenchCoding37.91%Overall accuracyVerifiedDec 18, 2025Details