BenchLeader
Menu
Best LLMsChoose a modelLeaderboardsModelsBenchmarksProvidersReleases9

Claude Opus 4.6 vs GPT-5.3-Codex

Full profiles: Claude Opus 4.6 (Anthropic) and GPT-5.3-Codex (OpenAI): every benchmark, price and rank.

Updated · Data from 9 sources

As of 26 Oct 2026, Claude Opus 4.6 (high) and GPT-5.3-Codex (xhigh) are neck and neck on quality (62.7 vs 63.8 on the BenchLeader Index, inside the ±5.5 margin of error), and GPT-5.3-Codex (xhigh) costs 2.1× less, which makes it the better buy for most workloads. Pick Claude Opus 4.6 for multimodal (67 vs 62); pick GPT-5.3-Codex for reasoning (69 vs 66).

Claude Opus 4.6 benchmarks high Anthropic GPT-5.3-Codex benchmarks xhigh OpenAIOpen in the comparison tool

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
CodingClaude Opus 4.663 vs 61 category score; WeirdML 78.0% vs 77.9%Medium
ReasoningGPT-5.3-Codex69 vs 66 category scoreMedium
Images and other mediaClaude Opus 4.667 vs 62 category scoreMedium
Chat quality (human preference)–Only Claude Opus 4.6 has an Arena rating (70 category score)Not comparable
Lowest costGPT-5.3-Codex$4.81 vs $10.00 per 1M blended; agentic coding session $0.161 vs $0.400–
Fastest first answerClaude Opus 4.61.2 s (Claude Opus 4.6 (high)) vs 55.6 s (GPT-5.3-Codex (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

  • Claude Opus 4.6 (high) and GPT-5.3-Codex (xhigh) are neck and neck on quality (62.7 vs 63.8, within the ±5.5 margin).
  • Claude Opus 4.6 (high) is stronger in coding, human preference, maths, multimodal.
  • GPT-5.3-Codex (xhigh) is stronger in reasoning, agents & tools, composite, instruction following, knowledge, long context.
  • GPT-5.3-Codex (xhigh) is 2.1× cheaper ($4.81 vs $10.00 per 1M blended).
  • GPT-5.3-Codex (xhigh) streams 2.7× faster (105 vs 39 tokens per second).

At a glance

SpecClaude Opus 4.6GPT-5.3-Codex
DeveloperAnthropicOpenAI
Released5 Feb 20265 Feb 2026
Context window (tokens)1,000,000400,000
Compared atClaude Opus 4.6 (high)GPT-5.3-Codex (xhigh)
Open weightsNoNo
Price in / out / cached input per 1M$5.00 / $25.00 / $0.500$1.75 / $14.00 / $0.175

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.

CategoryClaude Opus 4.6 (high)GPT-5.3-Codex (xhigh)Leader
Reasoning65.568.6GPT-5.3-Codex
Coding62.760.9Claude Opus 4.6
Agents & tools–80.1not comparable
Maths63.5–not comparable
Knowledge–66.4not comparable
Instruction following–73.5not comparable
Human preference70.0–not comparable
Multimodal66.661.5Claude Opus 4.6
Long context–66.8not comparable
Composite–65.8not comparable

Where sources disagree

Across the 1 benchmarks both have, Claude Opus 4.6 leads 1, GPT-5.3-Codex leads 0. Every overall index that rates both puts GPT-5.3-Codex ahead.

Overall indexClaude Opus 4.6 (high)GPT-5.3-Codex (xhigh)Leader
BenchLeader Index62.763.8neck and neck
AA Intelligence Index v4.3.226.4†32.5GPT-5.3-Codex

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; suite cost is what Artificial Analysis paid to run its whole Intelligence Index.

MeasureClaude Opus 4.6 (high)GPT-5.3-Codex (xhigh)
Blended price per 1M tokens$10.00$4.81
Input / output per 1M$5.00 / $25.00$1.75 / $14.00
Cached input per 1M$0.500$0.175
Price per task: chat reply$0.0095$0.0049
Price per task: summarize a 30-page report$0.075$0.029
Price per task: code edit$0.068$0.032
Price per task: agentic coding session$0.400$0.161
Price per task: structured extraction$0.015$0.0063
Output speed39 tok/s105 tok/s
Time to first answer1.2 s55.6 s
Time to a complete answer–60.4 s

Each model is compared at its best-ranked reasoning-effort setting. Claude Opus 4.6 (high) scores 62.7 and takes 1 s to its first answer token; Claude Opus 4.6 (thinking) scores 62.5 in 1 s, Claude Opus 4.6 scores 62.5 in 1 s. GPT-5.3-Codex (xhigh) scores 63.8 and takes 56 s to its first answer token. 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

111 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.

MetricPublisherClaude Opus 4.6 (high)GPT-5.3-Codex (xhigh)
BenchLeader IndexBenchLeader62.763.8
Coding scoreBenchLeader62.760.9
Human preference scoreBenchLeader70.0–
Maths scoreBenchLeader63.5–
Multimodal scoreBenchLeader66.661.5
Reasoning scoreBenchLeader65.568.6
Agents & tools scoreBenchLeader–80.1
Composite scoreBenchLeader–65.8
Instruction following scoreBenchLeader–73.5
Knowledge scoreBenchLeader–66.4
Long context scoreBenchLeader–66.8
Blended price $/MBenchLeader$10.00$4.81
Output speedBenchLeader39 tok/s105 tok/s
Time to first answerBenchLeader1.2 s55.6 s
Context windowBenchLeader1,000,000 tokens400,000 tokens
Reasoning
GPQA DiamondEpoch AI Benchmarking Hub90.5%†–
Humanity's Last ExamScale AI / CAIS19.0%†–
SimpleBenchSimpleBench67.6%†–
Arena Hard PromptsArena (LMArena)1530–
LiveBench 1.0 ReasoningLiveBench88.7%–
GPQA Diamond (AA)Artificial Analysis89.6%†91.5%
Humanity's Last Exam (AA)Artificial Analysis39.9%†42.5%
GPQA Diamond (Vals)Vals AI89.7%†–
Kagi LLM BenchmarkKagi LLM Benchmark72.4%†–
ARC-AGI-1ARC Prize94.0%–
ARC-AGI-2ARC Prize69.2%–
ARC-AGI-3ARC Prize0.5%†–
CritPtArtificial Analysis12.6%†16.9%
Chess PuzzlesEpoch AI Benchmarking Hub21.0%†–
Mystery Game PuzzlesEpoch AI Benchmarking Hub15.0%†–
CL-benchEpoch AI Benchmarking Hub20.7%†–
CL-bench LifeEpoch AI Benchmarking Hub17.0%–
LMCAEpoch AI Benchmarking Hub55.8%†–
DTBenchEpoch AI Benchmarking Hub91.2%†–
ForecastBenchForecastBench63.4%†–
Coding
SWE-bench Verified (Epoch)Epoch AI Benchmarking Hub78.7%†74.8%†
WeirdMLWeirdML78.0%77.9%
FrontierCodeCognition26.6%†–
GSO-BenchGSO-Bench41.2%–
Arena CodingArena (LMArena)1547–
Show all 111 benchmarks
MetricPublisherClaude Opus 4.6 (high)GPT-5.3-Codex (xhigh)
Coding
Arena WebDevArena (LMArena)15461407†
LiveBench 1.0 CodingLiveBench78.2%–
LiveCodeBenchVals AI84.7%†87.3%
IOIVals AI–53.8%
SWE-bench (Vals)Vals AI78.2%†78.0%
SWE-Bench ProScale AI SEAL51.9%†–
SWE-bench Verified (bash only)SWE-bench75.6%†–
SWE-bench Verified (any scaffold)SWE-bench75.6%†–
Vibe Code Bench v1.1Vals AI57.6%†61.8%
SWE-Bench Pro (private)Scale AI SEAL51.9%†–
SWE Atlas: Codebase QnAScale AI SEAL33.3%†–
SWE Atlas: RefactoringScale AI SEAL35.6%†–
SWE Atlas: Test WritingScale AI SEAL36.7%†–
ALE-BenchEpoch AI Benchmarking Hub996.5†1655.2
AlgoTuneAlgoTune1.5–
Agents & tools
Terminal-BenchTerminal-Bench79.8%†78.4%†
CybenchCybench93.0%†–
Remote Labor IndexScale AI / CAIS4.2%†–
APEX-AgentsMercor46.3%†–
LiveBench Agentic CodingLiveBench49.0%–
Terminal-Bench HardArtificial Analysis46.2%†53.0%
τ²-Bench Telecom (AA)Artificial Analysis92.1%†86.0%
MCP AtlasScale AI SEAL76.8%†–
HiL-BenchScale AI SEAL38.3%†4.3%†
τ³-bench Bankingτ²-bench27.3%†–
APEX-Agents (AA)Artificial Analysis33.0%†–
Terminal-Bench 2.0 (Vals)Vals AI58.4%†64.0%†
DeepResearch BenchDeepResearch Bench55.3%–
METR Time HorizonsMETR78.9%†74.5%†
GBAEvalEpoch AI Benchmarking Hub44.1%†–
Vending-Bench 2Andon Labs8017.6†5940.1†
Maths
FrontierMath Tiers 1–3Epoch AI Benchmarking Hub66.0%†–
FrontierMath Tier 4Epoch AI Benchmarking Hub26.8%†–
OTIS Mock AIMEEpoch AI Benchmarking Hub94.4%†–
ProofBenchVals AI50.0%†–
LiveBench MathematicsLiveBench89.3%–
AIME (Vals)Vals AI95.6%†–
AIME 2026MathArena96.7%–
HMMT February 2026MathArena96.2%–
USAMO 2026MathArena47.0%–
MathArena ApexMathArena34.5%–
Arena MathsArena (LMArena)1517–
Knowledge
SimpleQA VerifiedEpoch AI Benchmarking Hub47.0%†–
LiveBench 1.0 Data AnalysisLiveBench69.9%–
AA-OmniscienceArtificial Analysis13.7†10.9
MMLU-ProVals AI89.1%†–
LegalBenchVals AI85.3%†–
CorpFinVals AI67.0%†–
TaxEvalVals AI76.0%†–
MedQAVals AI95.4%†–
PRBench FinanceScale AI SEAL53.3%†–
PRBench LegalScale AI SEAL52.3%†–
MultiNRCScale AI SEAL57.1%†–
CaseLaw v2Vals AI62.1%†–
MedCodeVals AI48.2%†–
MedScribeVals AI86.7%†–
DrugDiscoveryBenchScale AI SEAL27.7%†–
EBR-benchEpoch AI Benchmarking Hub12.7%†–
AA-Omniscience: accuracyArtificial Analysis47.0%†52.9%
AA-Omniscience: non-hallucinationArtificial Analysis37.2%†10.8%
Global-MMLU-Lite (AA)Artificial Analysis92.2%†–
Instruction following
LiveBench LanguageLiveBench83.3%–
LiveBench Instruction FollowingLiveBench63.3%–
IFBenchArtificial Analysis53.1%†75.4%
MultiChallengeScale AI SEAL37.1%†–
TutorBenchScale AI SEAL53.7%†–
Arena Instruction FollowingArena (LMArena)1511–
Human preference
Arena TextArena (LMArena)1503–
EQ-Bench 4EQ-Bench1213†–
Arena Creative WritingArena (LMArena)1500–
Arena Multi-turnArena (LMArena)1516–
Arena Longer QueriesArena (LMArena)1521–
Multimodal
Arena VisionArena (LMArena)1316–
MMMU-ProArtificial Analysis75.4%†78.5%
VISTAScale AI SEAL46.1%†–
MMMU-Pro (Vals)Vals AI83.9%†–
MortgageTaxVals AI68.5%†–
SAGEVals AI51.6%†–
VTBScale AI SEAL27.5%†–
Long context
AA-LCRArtificial Analysis78.0%†83.3%
Arena DocumentArena (LMArena)1495–
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub155.7†157.1†
LiveBench 1.0LiveBench74.5%–
AA Intelligence Index v4.3.2Artificial Analysis31.9†32.5
Safety & honesty
FORTRESSScale AI SEAL13.0%†–
MASKScale AI SEAL85.4%†–

Data as of 2026-10-26. Best configuration of each model; every score links to its source on the model pages.

What changed

  • Claude Opus 4.6 (high): BenchLeader Index 60.4 → 62.9.
  • Claude Opus 4.6 (high): 1 new result (USAMO 2026).
  • GPT-5.3-Codex (xhigh): BenchLeader Index 64.4 → 63.9.
  • Claude Opus 4.6 (high): BenchLeader Index 61.1 → 60.4.
  • GPT-5.3-Codex (xhigh): BenchLeader Index 64.9 → 64.4.

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

Claude Opus 4.6 vs GPT-5.3-Codex: questions

Is Claude Opus 4.6 better than GPT-5.3-Codex?
Claude Opus 4.6 (high) and GPT-5.3-Codex (xhigh) are neck and neck on quality (62.7 vs 63.8, within the ±5.5 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 2026-10-26, so check the category scores for your use.
Which is cheaper, Claude Opus 4.6 or GPT-5.3-Codex, and will that change?
GPT-5.3-Codex is cheaper today: $4.81 against $10.00 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
Is Claude Opus 4.6 better than GPT-5.3-Codex for coding?
Claude Opus 4.6 scores higher in coding (63 vs 61 on the category index, where 50 is average). On WeirdML, the top shared coding benchmark, it is 78.0% for Claude Opus 4.6 against 77.9% for GPT-5.3-Codex.
Which is better for long documents?
Claude Opus 4.6 accepts more context: 1,000,000 against 400,000 tokens for GPT-5.3-Codex.
Which has lower latency?
Claude Opus 4.6 (high) answers first: 1.2 s against 55.6 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.
Which is faster, Claude Opus 4.6 or GPT-5.3-Codex?
GPT-5.3-Codex streams faster: 105 against 39 output tokens per second.
Do independent benchmarks agree on Claude Opus 4.6 vs GPT-5.3-Codex?
Across the 1 benchmarks both have, Claude Opus 4.6 leads 1, GPT-5.3-Codex leads 0. Every overall index that rates both puts GPT-5.3-Codex ahead.
When was Claude Opus 4.6 released?
Claude Opus 4.6 was released on 5 Feb 2026.