GPT-5.3 Codex

GPT-5.3 Codex is a reasoning model from OpenAI. 19 benchmarks count toward its score, in 6 categories.

availableShows if the model has enough results for an index.
IndexOverall score. 50 is the middle.64.8 ±5.2
CoverageShare of the index weight with results.85%
SpeedOutput tokens per second.80/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.
60.7
CodingCode writing and repair.
64.1
ReasoningLogic problems and puzzles.
69.8
MultimodalTasks with images and text.
63.0
KnowledgeFacts and expert knowledge.
68.8
MultilingualTasks in many languages.
N/A
InstructionTasks with strict rules in the prompt.
65.6
MathMath problems.
N/A

Results

19 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.
Artificial Analysis GPQA DiamondKnowledge91.5%62.3Artificial Analysis
LiveCodeBenchCoding87.3%63.71 Sept 2026Vals AI
τ²-Bench Tool-Agent-User EvaluationAgentic86.0%60.5Victor Barres et al.
Software Engineering Benchmark VerifiedCoding85.0%65.7Carlos E. Jimenez et al.
Artificial Analysis Long Context ReasoningReasoning83.3%65.9Artificial Analysis
Artificial Analysis MMMU-ProMultimodal78.5%63.0Artificial Analysis
SWE-benchCoding78.0%60.11 Sept 2026Vals AI
Artificial Analysis IFBenchInstruction75.4%65.6Artificial Analysis
SWE-Bench verifiedCoding74.8%57.5high effortEpoch AI
OSWorld-VerifiedAgentic64.7%53.9Tianbao Xie et al.
Terminal-Bench 2.0Agentic64.0%68.44 Jun 2026Vals AI
Vibe Code Bench v1.1Coding61.8%67.9OpenHands21 Sept 2026Vals AI
SWE-RebenchCoding58.2%Nebius
Gert Labs Composite Game BenchmarkAgentic57.5%63.5Gert Labs
SWE-bench ProCoding56.8%58.9Xiang Deng et al.
IOICoding53.8%69.121 Sept 2026Vals AI
Artificial Analysis Omniscience AccuracyKnowledge52.9%79.3Artificial Analysis
IOI v1Coding43.8%64.99 Aug 2026Vals AI
Artificial Analysis Humanity's Last ExamKnowledge42.5%70.6Artificial Analysis
JobBenchAgentic33.7%57.3Yuetai Li et al.
Artificial Analysis Intelligence IndexKnowledge32.5%63.0Artificial Analysis
Critical Physics TasksReasoning16.9%73.8Artificial Analysis

19 benchmarks count, from 21 of 22 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 statedVals AI, collected directlyNo licence stated. Read from the public leaderboard and credited to Vals AIEpoch AI, collected directlyCC BY — free to use and redistribute with attribution

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