BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases7

GPT-5.3-Codex vs GPT-5.5

Full profiles: GPT-5.3-Codex (OpenAI) and GPT-5.5 (OpenAI): every benchmark, price and rank.

Updated · Data from 13 sources

As of 6 Nov 2026, GPT-5.3-Codex and GPT-5.5 are neck and neck on quality (62.9 vs 64.3 on the BenchLeader Index, inside the ±7.9 margin), and GPT-5.3-Codex is 2.3× cheaper at $4.81 vs $11.25 per 1M tokens blended. Choose GPT-5.3-Codex for agents & tools (80 vs 62); choose GPT-5.5 for multimodal (62 vs 61).

Key facts (as of 6 Nov 2026)

  • GPT-5.3-Codex scores 62.9 on the BenchLeader Index and GPT-5.5 scores 64.3 (BenchLeader, 6 Nov 2026).
  • GPT-5.3-Codex costs $4.81 per 1M tokens blended (list price, 6 Nov 2026).
  • GPT-5.5 costs $11.25 per 1M tokens blended (list price, 6 Nov 2026).
  • Of the 15 benchmarks both models have, GPT-5.5 leads 14 (BenchLeader, 6 Nov 2026).
GPT-5.3-Codex benchmarks xhigh OpenAI GPT-5.5 benchmarks xhigh OpenAIAdd a third model

Which to pick for…

The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.

Use casePickWhyConfidence
CodingEither61 vs 61 category score; LiveCodeBench 87.3% vs 85.3%Either
Agents & tool useGPT-5.3-Codex80 vs 62 category score; Terminal-Bench Hard 53.0% vs 60.6%High
Long documentsEither67 vs 67 category score; AA-LCR v1.1 83.3% vs 84.3%Either
Images and other mediaGPT-5.562 vs 61 category score; MMMU-Pro 79.9% vs 78.5%Medium
Following instructionsEither74 vs 74 category score; IFBench 75.4% vs 75.8%Either
Lowest costGPT-5.3-Codex$4.81 vs $11.25 per 1M blended; agentic coding session $0.161 vs $0.420–
Fastest first answerEither55.5 s (GPT-5.3-Codex (xhigh)) vs 58.0 s (GPT-5.5 (xhigh)) to the first answer token–
High volume on a budgetNeither flagshipA smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs–

Key differences

  • GPT-5.3-Codex (xhigh) and GPT-5.5 (xhigh) are neck and neck on quality (62.9 vs 64.3, within the ±7.9 margin).
  • GPT-5.3-Codex (xhigh) is stronger in agents & tools.
  • GPT-5.5 (xhigh) is stronger in instruction following, long context, multimodal, knowledge, maths, reasoning.
  • GPT-5.3-Codex (xhigh) is 2.3× cheaper ($4.81 vs $11.25 per 1M blended).

At a glance

SpecGPT-5.3-CodexGPT-5.5
DeveloperOpenAIOpenAI
Released5 Feb 202623 Apr 2026
Context window (tokens)400,0001,050,000
Compared atGPT-5.3-Codex (xhigh)GPT-5.5 (xhigh)
Open weightsNoNo
Price in / out / cached input per 1M$1.75 / $14.00 / $0.175$5.00 / $30.00 / $0.500

Category by category

Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.

CategoryGPT-5.3-Codex (xhigh)GPT-5.5 (xhigh)Leader
Reasoning–68.9not comparable
Coding60.960.9level
Agents & tools80.061.7GPT-5.3-Codex
Maths–69.6not comparable
Knowledge–64.0not comparable
Instruction following73.573.9level
Multimodal60.962.4GPT-5.5
Long context66.567.1level

On one chart

Each domain on the index's own scale, where 50 is an average model. Click a domain to see the benchmarks behind it. Pick up to 5 models from the list.

2550 avg75CodingAgents & toolsInstructionfollowingLong contextMultimodalGPT-5.3-Codex, Coding: 60.9GPT-5.3-Codex, Agents & tools: 80.0GPT-5.3-Codex, Instruction following: 73.5GPT-5.3-Codex, Long context: 66.5GPT-5.3-Codex, Multimodal: 60.9GPT-5.5, Coding: 60.9GPT-5.5, Agents & tools: 61.7GPT-5.5, Instruction following: 73.9GPT-5.5, Long context: 67.1GPT-5.5, Multimodal: 62.4

Not drawn, measured for fewer than two of these models: Reasoning, Maths, Knowledge, Human preference.

ModelIndex

Where sources disagree

Across the 15 benchmarks both have, GPT-5.3-Codex leads 1, GPT-5.5 leads 14.

Overall indexGPT-5.3-Codex (xhigh)GPT-5.5 (xhigh)Leader
BenchLeader Index62.964.3neck and neck

Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).

Price, speed and latency

List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page. Cost, tokens and time per benchmark task are measured: Artificial Analysis’ averages over its whole Intelligence Index run.

MeasureGPT-5.3-Codex (xhigh)GPT-5.5 (xhigh)
Blended price per 1M tokens$4.81$11.25
Input / output per 1M$1.75 / $14.00$5.00 / $30.00
Cached input per 1M$0.175$0.500
Price per task: chat reply$0.0049$0.011
Price per task: summarize a 30-page report$0.029$0.078
Price per task: code edit$0.032$0.075
Price per task: agentic coding session$0.161$0.420
Price per task: structured extraction$0.0063$0.016
Agentic coding session, input mostly cached$0.090$0.217
Output speed74 tok/s79 tok/s
Time to first answer55.5 s58.0 s
Time to a complete answer111.8 s115.5 s

Each model is compared at its best-ranked reasoning-effort setting. GPT-5.3-Codex (xhigh) scores 62.9 and takes 55 s to its first answer token. GPT-5.5 (xhigh) scores 64.3 and takes 58 s to its first answer token; GPT-5.5 (high) scores 63.7 in 27 s, GPT-5.5 scores 61.7 in 2 s. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.

Every benchmark

126 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.

MetricPublisherGPT-5.3-Codex (xhigh)GPT-5.5 (xhigh)
BenchLeader IndexBenchLeader62.964.3
Agents & tools scoreBenchLeader80.061.7
Coding scoreBenchLeader60.960.9
Instruction following scoreBenchLeader73.573.9
Long context scoreBenchLeader66.567.1
Multimodal scoreBenchLeader60.962.4
Knowledge scoreBenchLeader–64.0
Maths scoreBenchLeader–69.6
Reasoning scoreBenchLeader–68.9
Blended price $/MBenchLeader$4.81$11.25
Output speedBenchLeader74 tok/s79 tok/s
Time to first answerBenchLeader55.5 s58.0 s
Context windowBenchLeader400,000 tokens1,050,000 tokens
Reasoning
GPQA DiamondEpoch AI Benchmarking Hub–90.7%†
SimpleBenchSimpleBench–69.0%†
Arena Hard PromptsArena (LMArena)–1499†
LiveBench 1.0 ReasoningLiveBench–89.7%
GPQA Diamond (AA)Artificial Analysis91.5%93.5%
Humanity's Last Exam (AA)Artificial Analysis42.5%45.8%
GPQA Diamond (Vals)Vals AI–93.2%
Kagi LLM BenchmarkKagi LLM Benchmark–88.8%†
ARC-AGI-1ARC Prize–95.0%
ARC-AGI-2ARC Prize–85.0%
ARC-AGI-3ARC Prize–0.4%†
CritPtArtificial Analysis16.9%27.1%
Chess PuzzlesEpoch AI Benchmarking Hub–63.0%
Mystery Game PuzzlesEpoch AI Benchmarking Hub–56.0%
CL-bench LifeEpoch AI Benchmarking Hub–22.2%†
LMCAEpoch AI Benchmarking Hub–54.3%
DTBenchEpoch AI Benchmarking Hub–96.0%
ForecastBenchForecastBench–64.5%†
BTF-3Epoch AI Benchmarking Hub–14.3%†
Coding
SWE-bench Verified (Epoch)Epoch AI Benchmarking Hub74.8%†–
SciCodeSciCode–55.8%
WeirdMLWeirdML77.9%84.9%
FrontierCodeCognition–43.0%†
GSO-BenchGSO-Bench–40.2%
Arena CodingArena (LMArena)–1510†
Show all 126 benchmarks
MetricPublisherGPT-5.3-Codex (xhigh)GPT-5.5 (xhigh)
Coding
Arena WebDevArena (LMArena)1407†1512
LiveBench 1.0 CodingLiveBench–82.2%
LiveCodeBenchVals AI87.3%85.3%
IOIVals AI53.8%–
SWE-bench (Vals)Vals AI78.0%82.6%
Code MigrationVals AI–45.2%
ProgramBenchVals AI–0.5%
Vibe Code Bench v1.1Vals AI61.8%69.8%
SWE Atlas: Codebase QnAScale AI SEAL–45.4%
SWE Atlas: RefactoringScale AI SEAL–44.8%
SWE Atlas: Test WritingScale AI SEAL–42.6%
MirrorCodeEpoch AI Benchmarking Hub–10.0%†
Surface Evolver BenchEpoch AI Benchmarking Hub–88.1%†
DeepSWE v1.1Epoch AI Benchmarking Hub–67.0%
ALE-BenchEpoch AI Benchmarking Hub1655.21943.0
Terminal-Bench 4.0 (AA)Artificial Analysis–14.7%
Terminal-Bench 2.1 (AA)Artificial Analysis–84.3%
GameBench 2 OverallDuelLab GameBench 2–49.0†
GameBench 2 by reasoning settingDuelLab GameBench 2–59.7
Agents & tools
Terminal-BenchTerminal-Bench78.4%†84.7%†
OSWorld-Verified 2.0OSWorld–13.0%
Remote Labor IndexScale AI / CAIS–6.3%†
APEX-AgentsMercor–55.1%†
LiveBench Agentic CodingLiveBench–54.0%
Terminal-Bench HardArtificial Analysis53.0%60.6%
τ²-Bench Telecom (AA)Artificial Analysis86.0%93.9%
Terminal-Bench 2.1 (Vals)Vals AI–76.4%†
MCP AtlasScale AI SEAL–75.3%
HiL-BenchScale AI SEAL4.3%†39.7%†
τ³-bench Bankingτ²-bench–44.6%
τ³-Banking (AA)Artificial Analysis–39.0%
ITBench SRE (AA)Artificial Analysis–45.8%
Analyst Agent (AA)Artificial Analysis–50.0%
APEX-Agents (AA)Artificial Analysis–37.7%
Finance Agent v2Vals AI–51.8%
Harvey's Legal Agent Benchmark v1Vals AI–3.8%
Legal Research BenchVals AI–40.4%
SkillsBenchVals AI–62.2%
SREBenchVals AI–3.8%
Tax Agent BenchVals AI–21.8%
Terminal-Bench 2.0 (Vals)Vals AI64.0%†73.2%†
Time Horizon Index: KSPVals AI–8.3%
DeepResearch BenchDeepResearch Bench–54.0%†
PostTrainBenchEpoch AI Benchmarking Hub–27.2%
ExploitBenchEpoch AI Benchmarking Hub–47.4%†
METR Time HorizonsMETR74.5%†–
GBAEvalEpoch AI Benchmarking Hub–53.2%†
Vending-Bench 2Andon Labs5940.1†7523.8†
AutomationBenchArtificial Analysis–47.3%
GDP.pdf (AA)Artificial Analysis–21.2%
EnterpriseOps-GymArtificial Analysis–46.6%
AA-Briefcase v1.1Artificial Analysis8681135
Maths
FrontierMath Tiers 1–3Epoch AI Benchmarking Hub–85.3%
FrontierMath Tier 4Epoch AI Benchmarking Hub–72.5%
OTIS Mock AIMEEpoch AI Benchmarking Hub–84.4%†
ProofBenchVals AI–50.0%
LiveBench MathematicsLiveBench–95.9%
AIME 2026MathArena–100.0%
HMMT February 2026MathArena–98.5%
USAMO 2026MathArena–98.2%
MathArena ApexMathArena–80.2%
Arena MathsArena (LMArena)–1500†
Knowledge
SimpleQA VerifiedEpoch AI Benchmarking Hub–63.0%
LiveBench 1.0 Data AnalysisLiveBench–81.6%
MMLU-ProVals AI–88.1%
LegalBenchVals AI–86.5%
CorpFinVals AI–68.4%
TaxEvalVals AI–75.0%
CaseLaw v2Vals AI–66.2%
Excel Modeling BenchmarkVals AI–64.5%
MedCodeVals AI–49.1%
MedScribeVals AI–86.9%
Public Benefits Bench v1.1Vals AI–60.9%
DrugDiscoveryBenchScale AI SEAL–51.6%
EBR-benchEpoch AI Benchmarking Hub–34.3%
GDP.pdfEpoch AI Benchmarking Hub–26.0%
Instruction following
LiveBench LanguageLiveBench–87.4%
LiveBench Instruction FollowingLiveBench–70.7%
IFBenchArtificial Analysis75.4%75.8%
Arena Instruction FollowingArena (LMArena)–1475†
Human preference
Arena TextArena (LMArena)–1477†
EQ-Bench 4EQ-Bench–1295†
Arena Creative WritingArena (LMArena)–1448†
Arena Multi-turnArena (LMArena)–1484†
Arena Longer QueriesArena (LMArena)–1487†
Multimodal
Arena VisionArena (LMArena)–1297†
MMMU-ProArtificial Analysis78.5%79.9%
MMMU-Pro (Vals)Vals AI–88.3%
MortgageTaxVals AI–68.8%
Blueprint-Bench 2Epoch AI Benchmarking Hub–36.2%†
Code2Video-AAArtificial Analysis–1000
Long context
AA-LCR v1.1Artificial Analysis83.3%84.3%
Arena DocumentArena (LMArena)–1483†
MLCRArtificial Analysis–28.9%
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub–160.0†
LiveBench 1.0LiveBench–80.2%
AA Intelligence Index v4.3.2Artificial Analysis32.538.4
AA Progress Index v1Artificial Analysis301.3320
Vals RSI IndexVals AI–16.5%
Vals Multimodal IndexVals AI–68.1%
Safety & honesty
FORTRESSScale AI SEAL–16.3%

Data as of 2026-11-06. Best configuration of each model; every score links to its source on the model pages.

DuelLab GameBench 2, release gb2-20261103-09b2218cbd85 (https://benchmarks.duellab.org/releases/gb2-20261103-09b2218cbd85/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.

What changed

  • GPT-5.5 (xhigh): 1 score revision.
  • GPT-5.3-Codex (xhigh): 1 new result (AA-Briefcase v1.1).
  • GPT-5.3-Codex (xhigh): 1 new result (AA Progress Index v1).
  • GPT-5.5 (xhigh): 1 new result (AA Progress Index v1).
  • GPT-5.5 (xhigh): cost per task moved from $2.63/M to $2.10/M (20% down).
  • GPT-5.5 (xhigh): 1 new result (AA-Briefcase v1.1).
  • GPT-5.5 (xhigh): 2 new results (Harvey's Legal Agent Benchmark v1, Vals RSI Index).
  • GPT-5.5 (xhigh): 1 score revision.
  • GPT-5.5 (xhigh): BenchLeader Index 65.1 → 64.6.
  • GPT-5.5 (xhigh): 6 new results (USAMO 2026, AutomationBench, GDP.pdf (AA)…).

Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs

GPT-5.3-Codex vs GPT-5.5: questions

Is GPT-5.3-Codex better than GPT-5.5?
GPT-5.3-Codex (xhigh) and GPT-5.5 (xhigh) are neck and neck on quality (62.9 vs 64.3, within the ±7.9 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 6 Nov 2026, so check the category scores for your use.
GPT-5.3-Codex vs GPT-5.5: which is cheaper, and will that change?
GPT-5.3-Codex is cheaper today: $4.81 against $11.25 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
Is GPT-5.3-Codex or GPT-5.5 better for coding?
They are level in coding (61 vs 61 on the category index). On LiveCodeBench, the top shared coding benchmark, it is 87.3% for GPT-5.3-Codex against 85.3% for GPT-5.5.
Which is faster, GPT-5.3-Codex or GPT-5.5?
GPT-5.5 streams faster (79 vs 74 output tokens per second, Artificial Analysis 7-day median).
Is GPT-5.3-Codex or GPT-5.5 better for agents and tool use?
GPT-5.3-Codex scores higher in agentic tasks (80 vs 62 on the category index, where 50 is average). On Terminal-Bench Hard, the top shared agents & tools benchmark, it is 53.0% for GPT-5.3-Codex against 60.6% for GPT-5.5.
More questions (4)
Which is better for long documents?
They are level on long-context benchmarks (67 vs 67). GPT-5.5 accepts more context: 1,050,000 against 400,000 tokens for GPT-5.3-Codex.
Which has lower latency?
GPT-5.3-Codex answers first: 55.5 s against 58.0 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
Do independent benchmarks agree on GPT-5.3-Codex vs GPT-5.5?
Across the 15 benchmarks both have, GPT-5.3-Codex leads 1, GPT-5.5 leads 14.
When was GPT-5.5 released?
GPT-5.5 was released on 23 Apr 2026.