GPT-5.2-Codex

GPT-5.2-Codex is a reasoning model from OpenAI. 20 benchmarks count toward its score, in 7 categories.

availableShows if the model has enough results for an index.
IndexOverall score. 50 is the middle.62.1 ±4.1
CoverageShare of the index weight with results.95%
SpeedOutput tokens per second.15/s
Input / 1MUS dollars per 1M input tokens.$1.75
Output / 1MUS dollars per 1M output tokens.$14
ContextMaximum tokens in one request.400K
EloLMArena rating and rank.N/A

50 is the middle of the board. The range shows the doubt in the index.

CapabilitiesScore per category. 50 is the middle.

50 is the middle
AgenticMulti-step tasks with tools.
59.6
CodingCode writing and repair.
63.6
ReasoningLogic problems and puzzles.
62.0
MultimodalTasks with images and text.
60.3
KnowledgeFacts and expert knowledge.
61.2
MultilingualTasks in many languages.
N/A
InstructionTasks with strict rules in the prompt.
66.4
MathMath problems.
66.5

Results

20 counted
BenchmarkThe test name.CategoryThe capability that the test measures.ResultThe score from the publisher.IndexThis result on the index scale.RunThe settings of the run.DateDate of the result.Published byThe source of the result.
τ²-Bench Tool-Agent-User EvaluationAgentic92.1%64.9Victor Barres et al.
Artificial Analysis GPQA DiamondKnowledge89.9%60.7Artificial Analysis
LiveBench MathematicsMath88.8%66.525 Jun 2026LiveBench
LiveCodeBenchCoding88.0%64.31 Sept 2026Vals AI
LiveBench CodingCoding83.6%76.425 Jun 2026LiveBench
Artificial Analysis Long Context ReasoningReasoning82.3%65.2Artificial Analysis
LiveBench Data AnalysisReasoning78.2%64.625 Jun 2026LiveBench
LiveBench ReasoningReasoning77.7%63.925 Jun 2026LiveBench
Artificial Analysis IFBenchInstruction77.6%67.8Artificial Analysis
Artificial Analysis MMMU-ProMultimodal76.3%60.3Artificial Analysis
LiveBench LanguageKnowledge73.7%59.525 Jun 2026LiveBench
SWE-bench VerifiedCoding72.8%55.9mini-SWE-agent1 Sept 2026SWE-bench team
SWE-benchCoding72.4%55.61 Sept 2026Vals AI
LiveBench Instruction FollowingInstruction66.4%65.025 Jun 2026LiveBench
SWE-bench MultilingualCoding66.3%mini-SWE-agent2 Sept 2026SWE-bench team
Gert Labs Composite Game BenchmarkAgentic51.8%58.5Gert Labs
LiveBench Agentic CodingAgentic49.4%63.225 Jun 2026LiveBench
Artificial Analysis Omniscience AccuracyKnowledge41.1%64.7Artificial Analysis
Vibe Code Bench v1.1Coding37.9%57.9OpenHands21 Sept 2026Vals AI
Artificial Analysis Humanity's Last ExamKnowledge35.7%63.3Artificial Analysis
Artificial Analysis Intelligence IndexKnowledge28.5%58.0Artificial Analysis
JobBenchAgentic26.0%52.0Yuetai Li et al.
Critical Physics TasksReasoning8.7%56.6Artificial Analysis

20 benchmarks count, from 22 of 23 results. A grey row does not count. Too few models took that benchmark.

Sources

BenchLM benchmark aggregationUsed with attribution; per-benchmark results credited to their original authorsOpenRouter, collected directlyNo licence statedLiveBench, collected directlyNo licence stated for the leaderboard. The site repo serving these CSVs has no LICENSE; the harness repo carries upstream Apache-2.0 and MIT copies that cover code, not resultsVals AI, collected directlyNo licence stated. Read from the public leaderboard and credited to Vals AISWE-bench team, collected directlyNo licence stated. The repository publishes submission records for reproducibility and transparency and asks that SWE-bench be cited

Same level, lower price

Qwen3.8-Flash-Next66.4 · Freedots3-note Preview65.5 · FreeDeepSeek V4 Flash 042363.0 · $0.177

More from OpenAI

GPT-6 Astra83.5GPT-5.6 Sol76.7GPT-5.6 Terra73.4GPT-6 Sol76.9GPT-5.572.3