BenchLeader
Menu
Best LLMsChoose a modelLeaderboardsModelsBenchmarksProvidersReleases9

Claude Opus 5 vs GPT-5.5

Full profiles: Claude Opus 5 (Anthropic) and GPT-5.5 (OpenAI): every benchmark, price and rank.

Updated · Data from 12 sources

As of 25 Oct 2026, Claude Opus 5 (max) and GPT-5.5 (xhigh) are neck and neck on quality (68.8 vs 66.1, inside the ±4.6 margin of error) and cost about the same, so speed decides it: GPT-5.5 (xhigh) streams 1.5× faster. Pick Claude Opus 5 for coding (70 vs 61); pick GPT-5.5 for long context (67 vs 65).

Claude Opus 5 benchmarks max Anthropic GPT-5.5 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 570 vs 61 category score; Arena WebDev 1691 vs 1513High
Agents & tool useClaude Opus 564 vs 62 category score; τ³-Banking (AA) 42.1% vs 39.0%Medium
MathsEither71 vs 70 category score; FrontierMath Tiers 1–3 85.6% vs 85.3%Either
ReasoningGPT-5.576 vs 74 category score; ARC-AGI-2 85.0% vs 90.4%Medium
KnowledgeClaude Opus 572 vs 65 category score; AA-Omniscience 37.1 vs 20.5Medium
Long documentsGPT-5.567 vs 65 category score; AA-LCR 84.3% vs 79.3%Medium
Images and other mediaClaude Opus 568 vs 63 category score; MMMU-Pro 84.7% vs 79.9%Medium
Chat quality (human preference)–Only Claude Opus 5 has an Arena rating (69 category score)Not comparable
Lowest costEither$10.00 vs $11.25 per 1M blended; agentic coding session $0.400 vs $0.420–
Fastest first answerClaude Opus 524.1 s (Claude Opus 5 (max)) vs 39.1 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

  • Claude Opus 5 (max) and GPT-5.5 (xhigh) are neck and neck on quality (68.8 vs 66.1, within the ±4.6 margin).
  • Claude Opus 5 (max) is stronger in agents & tools, coding, composite, human preference, knowledge, maths, multimodal.
  • GPT-5.5 (xhigh) is stronger in long context, reasoning, instruction following.
  • They cost about the same ($10.00 per 1M blended).
  • GPT-5.5 (xhigh) streams 1.5× faster (86 vs 57 tokens per second).

At a glance

SpecClaude Opus 5GPT-5.5
DeveloperAnthropicOpenAI
Released24 Jul 202623 Apr 2026
Context window (tokens)1,000,0001,050,000
Compared atClaude Opus 5 (max)GPT-5.5 (xhigh)
Open weightsNoNo
Price in / out / cached input per 1M$5.00 / $25.00 / $0.500$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.

CategoryClaude Opus 5 (max)GPT-5.5 (xhigh)Leader
Reasoning74.175.5GPT-5.5
Coding70.261.1Claude Opus 5
Agents & tools64.362.3Claude Opus 5
Maths70.669.7level
Knowledge71.665.3Claude Opus 5
Instruction following–73.9not comparable
Human preference68.5–not comparable
Multimodal67.962.9Claude Opus 5
Long context64.767.3GPT-5.5
Composite86.372.4Claude Opus 5

Where sources disagree

Across the 45 benchmarks both have, Claude Opus 5 leads 34, GPT-5.5 leads 11. Every overall index that rates both puts Claude Opus 5 ahead.

Overall indexClaude Opus 5 (max)GPT-5.5 (xhigh)Leader
BenchLeader Index68.866.1neck and neck
AA Intelligence Index v4.3.250.838.4Claude Opus 5
Arena Text14911482†Claude Opus 5

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 5 (max)GPT-5.5 (xhigh)
Blended price per 1M tokens$10.00$11.25
Input / output per 1M$5.00 / $25.00$5.00 / $30.00
Cached input per 1M$0.500$0.500
Price per task: chat reply$0.0095$0.011
Price per task: summarize a 30-page report$0.075$0.078
Price per task: code edit$0.068$0.075
Price per task: agentic coding session$0.400$0.420
Price per task: structured extraction$0.015$0.016
Cost to run a full benchmark suite$5.86$2.63
Output speed57 tok/s86 tok/s
Time to first answer24.1 s39.1 s
Time to a complete answer32.9 s45.0 s

Per token is not per task: on Artificial Analysis' full suite Claude Opus 5 (max) cost $5.86 and GPT-5.5 (xhigh) $2.63, because Claude Opus 5 (max) used about 2.5× as many tokens for the same work, which a per-token price hides.

Each model is compared at its best-ranked reasoning-effort setting. Claude Opus 5 (max) scores 68.8 and takes 24 s to its first answer token; Claude Opus 5 (high) scores 67.9 in 10 s, Claude Opus 5 (xhigh) scores 67.4 in 25 s. GPT-5.5 (xhigh) scores 66.1 and takes 39 s to its first answer token; GPT-5.5 (high) scores 65.3 in 14 s, GPT-5.5 (medium) scores 62.9 in 3 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

149 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 5 (max)GPT-5.5 (xhigh)
BenchLeader IndexBenchLeader68.866.1
Agents & tools scoreBenchLeader64.362.3
Coding scoreBenchLeader70.261.1
Composite scoreBenchLeader86.372.4
Human preference scoreBenchLeader68.5–
Knowledge scoreBenchLeader71.665.3
Long context scoreBenchLeader64.767.3
Maths scoreBenchLeader70.669.7
Multimodal scoreBenchLeader67.962.9
Reasoning scoreBenchLeader74.175.5
Instruction following scoreBenchLeader–73.9
Blended price $/MBenchLeader$10.00$11.25
Output speedBenchLeader57 tok/s86 tok/s
Time to first answerBenchLeader24.1 s39.1 s
Context windowBenchLeader1,000,000 tokens1,050,000 tokens
Reasoning
GPQA DiamondEpoch AI Benchmarking Hub93.9%90.7%†
SimpleBenchSimpleBench80.6%†69.0%†
Arena Hard PromptsArena (LMArena)15151499†
LiveBench 2.0 ReasoningLiveBench91.2%†89.7%†
LiveBench 1.0 ReasoningLiveBench91.2%89.7%
GPQA Diamond (AA)Artificial Analysis93.2%93.5%
Humanity's Last Exam (AA)Artificial Analysis54.9%45.8%
GPQA Diamond (Vals)Vals AI93.4%†93.2%
Kagi LLM BenchmarkKagi LLM Benchmark–88.8%†
ARC-AGI-1ARC Prize97.5%95.0%
ARC-AGI-2ARC Prize90.4%85.0%
ARC-AGI-3ARC Prize30.2%†0.4%†
CritPtArtificial Analysis29.1%27.1%
MysteryMechanismVals AI37.4%†–
Chess PuzzlesEpoch AI Benchmarking Hub43.0%63.0%
Mystery Game PuzzlesEpoch AI Benchmarking Hub59.0%56.0%
CL-bench LifeEpoch AI Benchmarking Hub–22.2%†
LMCAEpoch AI Benchmarking Hub63.3%54.3%
DTBenchEpoch AI Benchmarking Hub97.6%96.0%
ForecastBenchForecastBench66.5%64.4%†
BTF-3Epoch AI Benchmarking Hub12.0%†14.3%†
Coding
SciCodeSciCode56.4%55.8%
WeirdMLWeirdML91.8%84.9%
FrontierCodeCognition53.4%43.0%†
GSO-BenchGSO-Bench–40.2%
Show all 149 benchmarks
MetricPublisherClaude Opus 5 (max)GPT-5.5 (xhigh)
Coding
Arena CodingArena (LMArena)15281511†
Arena WebDevArena (LMArena)16911513
LiveBench 2.0 CodingLiveBench90.8%†85.1%†
LiveBench 1.0 CodingLiveBench81.5%82.2%
LiveCodeBenchVals AI89.0%†85.3%
IOIVals AI84.3%†–
SWE-bench (Vals)Vals AI97.0%†82.6%
Code MigrationVals AI57.5%†45.2%
ProgramBenchVals AI3.0%†0.5%
Vibe Code Bench 1-100Vals AI28.5%†–
Vibe Code Bench v1.1Vals AI88.4%†69.8%
SWE Atlas: Codebase QnAScale AI SEAL63.2%†45.4%
SWE Atlas: RefactoringScale AI SEAL–44.8%
SWE Atlas: Test WritingScale AI SEAL62.2%†42.6%
MirrorCodeEpoch AI Benchmarking Hub–10.0%†
Surface Evolver BenchEpoch AI Benchmarking Hub–88.1%†
DeepSWE v1.1Epoch AI Benchmarking Hub73.7%67.0%
CursorBenchCursor46.6%–
ALE-BenchEpoch AI Benchmarking Hub2164.6†1943.0
FrontierSWEEpoch AI Benchmarking Hub52.0%–
Terminal-Bench 4.0 (AA)Artificial Analysis49.0%14.7%
Terminal-Bench 2.1 (AA)Artificial Analysis89.1%84.3%
Startrise FrontendStartrise82.3%†–
Agents & tools
Terminal-BenchTerminal-Bench51.8%84.7%†
OSWorld-Verified 2.0OSWorld31.4%13.0%
Remote Labor IndexScale AI / CAIS–6.3%†
APEX-AgentsMercor65.8%55.1%†
Arena AgentArena (LMArena)8.2–
LiveBench Long AgenticLiveBench62.1%†44.6%†
LiveBench BrowserLiveBench84.0%†48.0%†
LiveBench Agentic CodingLiveBench65.2%54.0%
Terminal-Bench HardArtificial Analysis–60.6%
τ²-Bench Telecom (AA)Artificial Analysis–93.9%
Terminal-Bench 2.1 (Vals)Vals AI84.6%†76.4%†
MCP AtlasScale AI SEAL85.8%†75.3%
HiL-BenchScale AI SEAL57.0%†39.7%†
τ³-bench Bankingτ²-bench48.7%44.6%
τ³-Banking (AA)Artificial Analysis42.1%39.0%
ITBench SRE (AA)Artificial Analysis–45.8%
Analyst Agent (AA)Artificial Analysis53.8%50.0%
APEX-Agents (AA)Artificial Analysis–37.7%
CUA-benchVals AI9.0%–
CyberBenchVals AI65.4%†–
Finance Agent v2Vals AI58.6%†51.8%
Harvey's Legal Agent Benchmark v1Vals AI6.7%†3.8%
Legal Research BenchVals AI55.3%†40.4%
SkillsBenchVals AI60.4%†62.2%
SREBenchVals AI12.2%†3.8%
Tax Agent BenchVals AI45.5%†21.8%
Terminal-Bench 2.0 (Vals)Vals AI–73.2%†
Terminal-Bench 4.0 (Vals)Vals AI53.5%†–
Terminal-Bench ScienceVals AI27.1%†–
Time Horizon Index: KSPVals AI18.8%†8.3%
Arena SearchArena (LMArena)1214†–
BALROGBALROG63.4%–
DeepResearch BenchDeepResearch Bench–54.0%†
PostTrainBenchEpoch AI Benchmarking Hub35.0%†27.2%
ExploitBenchEpoch AI Benchmarking Hub–47.4%†
GBAEvalEpoch AI Benchmarking Hub79.6%†53.2%†
Vending-Bench 2Andon Labs11181.9†7523.8†
AutomationBenchArtificial Analysis56.6%47.3%
GDP.pdf (AA)Artificial Analysis21.6%21.2%
EnterpriseOps-GymArtificial Analysis47.5%46.6%
AA-Briefcase v1.1Artificial Analysis16571136
Terminal-Bench-Science 0.1 (AA)Artificial Analysis28.6%–
DeepSec-Bench (AA)Artificial Analysis25.9%–
FACTS SearchGoogle85.3%†–
Maths
FrontierMath Tiers 1–3Epoch AI Benchmarking Hub85.6%85.3%
FrontierMath Tier 4Epoch AI Benchmarking Hub73.2%72.5%
OTIS Mock AIMEEpoch AI Benchmarking Hub98.9%84.4%†
ProofBenchVals AI99.0%50.0%
LiveBench MathematicsLiveBench95.7%95.9%
AIME 2026MathArena–100.0%
HMMT February 2026MathArena–98.5%
USAMO 2026MathArena–98.2%
MathArena ApexMathArena–80.2%
Arena MathsArena (LMArena)15171501†
Knowledge
SimpleQA VerifiedEpoch AI Benchmarking Hub59.9%63.0%
LiveBench 2.0 Data AnalysisLiveBench74.6%†81.6%†
LiveBench ResearchLiveBench88.9%†61.6%†
LiveBench 1.0 Data AnalysisLiveBench74.5%81.6%
AA-OmniscienceArtificial Analysis37.120.5
MMLU-ProVals AI91.6%†88.1%
LegalBenchVals AI87.0%†86.5%
CorpFinVals AI73.2%†68.4%
TaxEvalVals AI75.1%†75.0%
BioMysteryBenchVals AI79.3%†–
CaseLaw v2Vals AI–66.2%
Excel Modeling BenchmarkVals AI73.6%†64.5%
MedCodeVals AI63.6%†49.1%
MedScribeVals AI91.0%†86.9%
Public Benefits Bench v1.1Vals AI76.9%†60.9%
DrugDiscoveryBenchScale AI SEAL53.7%51.6%
EBR-benchEpoch AI Benchmarking Hub45.7%34.3%
GDP.pdfEpoch AI Benchmarking Hub24.0%26.0%
AA-Omniscience: accuracyArtificial Analysis60.9%58.0%
AA-Omniscience: non-hallucinationArtificial Analysis39.2%11.0%
Instruction following
LiveBench LanguageLiveBench88.7%87.4%
LiveBench Instruction FollowingLiveBench63.8%70.7%
IFBenchArtificial Analysis–75.8%
Arena Instruction FollowingArena (LMArena)14951475†
Human preference
Arena TextArena (LMArena)14911477†
EQ-Bench 4EQ-Bench1360†1294†
Arena Creative WritingArena (LMArena)14741448†
Arena Multi-turnArena (LMArena)14871484†
Arena Longer QueriesArena (LMArena)15031487†
Multimodal
Arena VisionArena (LMArena)1320†1297†
LiveBench DesignLiveBench83.0%†70.3%†
MMMU-ProArtificial Analysis84.7%79.9%
MMMU-Pro (Vals)Vals AI89.9%†88.3%
MortgageTaxVals AI72.1%†68.8%
SAGEVals AI49.4%†51.5%
Blueprint-Bench 2Epoch AI Benchmarking Hub30.4%†36.2%†
Long context
AA-LCRArtificial Analysis79.3%84.3%
Arena DocumentArena (LMArena)1490†1483†
MLCRArtificial Analysis11.1%28.9%
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub163.9†159.7†
LiveBench 2.0LiveBench82.1%†68.7%†
LiveBench 1.0LiveBench80.1%80.2%
AA Intelligence Index v4.3.2Artificial Analysis50.838.4
Vals RSI IndexVals AI33.0%16.5%
Vals Science IndexVals AI47.9%†–
Vals Multimodal IndexVals AI73.9%†68.1%
Safety & honesty
FORTRESSScale AI SEAL16.8%†16.3%

Data as of 2026-10-25. Best configuration of each model; every score links to its source on the model pages.

What changed

  • GPT-5.5 (xhigh): 1 new result (AA-Briefcase v1.1).
  • Claude Opus 5 (max): 1 new result (Vals RSI Index).
  • GPT-5.5 (xhigh): 2 new results (Harvey's Legal Agent Benchmark v1, Vals RSI Index).
  • Claude Opus 5 (max): 1 score revision.
  • GPT-5.5 (xhigh): 1 score revision.
  • Claude Opus 5 (max): 1 new result (ForecastBench); 1 score revision.
  • Claude Opus 5 (max): BenchLeader Index 68.3 → 68.9.
  • Claude Opus 5 (max): 4 new results (GDPval-AA v2.1, AA-Omniscience: accuracy, AA-Omniscience: non-hallucination…).
  • GPT-5.5 (xhigh): 6 new results (USAMO 2026, AutomationBench, GDP.pdf (AA)…).
  • Claude Opus 5 (max): 1 score revision.

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

Claude Opus 5 vs GPT-5.5: questions

Is Claude Opus 5 better than GPT-5.5?
Claude Opus 5 (max) and GPT-5.5 (xhigh) are neck and neck on quality (68.8 vs 66.1, within the ±4.6 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 2026-10-25, so check the category scores for your use.
Which is cheaper, Claude Opus 5 or GPT-5.5, and will that change?
Claude Opus 5 is cheaper today: $10.00 against $11.25 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
Is Claude Opus 5 better than GPT-5.5 for coding?
Claude Opus 5 scores higher in coding (70 vs 61 on the category index, where 50 is average). On Arena WebDev, the top shared coding benchmark, it is 1691 for Claude Opus 5 against 1513 for GPT-5.5.
Is Claude Opus 5 better than GPT-5.5 for agents and tool use?
Claude Opus 5 scores higher in agentic tasks (64 vs 62 on the category index, where 50 is average). On τ³-Banking (AA), the top shared agents & tools benchmark, it is 42.1% for Claude Opus 5 against 39.0% for GPT-5.5.
Which is better for long documents?
GPT-5.5 scores higher on long-context benchmarks (67 vs 65). Both take about 1.1M tokens: Claude Opus 5 1,000,000, GPT-5.5 1,050,000.
Which has lower latency?
Claude Opus 5 (max) answers first: 24.1 s against 39.1 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 5 or GPT-5.5?
GPT-5.5 streams faster: 86 against 57 output tokens per second.
Do independent benchmarks agree on Claude Opus 5 vs GPT-5.5?
Across the 45 benchmarks both have, Claude Opus 5 leads 34, GPT-5.5 leads 11. Every overall index that rates both puts Claude Opus 5 ahead.
When was Claude Opus 5 released?
Claude Opus 5 was released on 24 Jul 2026.