BenchLeader
Menu
Best LLMsChoose a modelLeaderboardsModelsBenchmarksProvidersReleases9

Claude Opus 5 vs o3

Full profiles: Claude Opus 5 (Anthropic) and o3 (OpenAI): every benchmark, price and rank.

Updated · Data from 18 sources

As of 26 Oct 2026, Claude Opus 5 (max) leads on quality (68.8 vs 59.5 on the BenchLeader Index), but o3 costs 2.9× less, so it is the better buy unless you need the extra quality. Pick Claude Opus 5 for reasoning (74 vs 52); pick o3 for agents & tools (70 vs 64).

Claude Opus 5 benchmarks max Anthropic o3 benchmarks 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 62 category score; Arena Coding 1528 vs 1460High
Agents & tool useo370 vs 64 category scoreMedium
MathsEither71 vs 71 category score; Arena Maths 1517 vs 1448Either
ReasoningClaude Opus 574 vs 52 category score; CritPt 29.1% vs 1.1%High
KnowledgeClaude Opus 572 vs 55 category score; AA-Omniscience 37.1 vs -15.6High
Long documentsClaude Opus 565 vs 62 category score; AA-LCR 79.3% vs 74.7%Medium
Images and other mediaClaude Opus 568 vs 57 category score; MMMU-Pro 84.7% vs 70.1%High
Chat quality (human preference)Claude Opus 569 vs 61 on Arena votesMedium
Lowest costo3$3.50 vs $10.00 per 1M blended; agentic coding session $0.152 vs $0.400–
Fastest first answero36.9 s (o3) vs 68.7 s (Claude Opus 5 (max)) 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) leads on quality: 68.8 vs 59.5.
  • Claude Opus 5 (max) is stronger in coding, composite, human preference, knowledge, long context, multimodal, reasoning.
  • o3 is stronger in agents & tools, instruction following.
  • o3 is 2.9× cheaper ($3.50 vs $10.00 per 1M blended).
  • o3 streams 3.2× faster (166 vs 51 tokens per second).

At a glance

SpecClaude Opus 5o3
DeveloperAnthropicOpenAI
Released24 Jul 202616 Apr 2025
Context window (tokens)1,000,000200,000
Compared atClaude Opus 5 (max)o3
Open weightsNoNo
Price in / out / cached input per 1M$5.00 / $25.00 / $0.500$2.00 / $8.00 / $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)o3Leader
Reasoning74.151.9Claude Opus 5
Coding70.261.8Claude Opus 5
Agents & tools64.370.3o3
Maths70.670.6level
Knowledge71.655.3Claude Opus 5
Instruction following–65.1not comparable
Human preference68.561.3Claude Opus 5
Multimodal67.956.9Claude Opus 5
Long context64.762.3Claude Opus 5
Composite86.352.0Claude Opus 5

Where sources disagree

Across the 16 benchmarks both have, Claude Opus 5 leads 16, o3 leads 0. Every overall index that rates both puts Claude Opus 5 ahead.

Overall indexClaude Opus 5 (max)o3Leader
BenchLeader Index68.859.5Claude Opus 5
AA Intelligence Index v4.3.250.820.2Claude Opus 5
Arena Text14911432Claude 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)o3
Blended price per 1M tokens$10.00$3.50
Input / output per 1M$5.00 / $25.00$2.00 / $8.00
Cached input per 1M$0.500$0.500
Price per task: chat reply$0.0095$0.0032
Price per task: summarize a 30-page report$0.075$0.029
Price per task: code edit$0.068$0.024
Price per task: agentic coding session$0.400$0.152
Price per task: structured extraction$0.015$0.0056
Cost to run a full benchmark suite$5.86–
Output speed51 tok/s166 tok/s
Time to first answer68.7 s6.9 s
Time to a complete answer78.4 s9.9 s

Each model is compared at its best-ranked reasoning-effort setting. Claude Opus 5 (max) scores 68.8 and takes 69 s to its first answer token; Claude Opus 5 (high) scores 67.9 in 17 s, Claude Opus 5 (xhigh) scores 67.4 in 31 s. o3 scores 59.5 and takes 7 s to its first answer token; o3 (high) scores 53.9 in 7 s, o3 (medium) scores 53.3 in 7 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

166 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)o3
BenchLeader IndexBenchLeader68.859.5
Agents & tools scoreBenchLeader64.370.3
Coding scoreBenchLeader70.261.8
Composite scoreBenchLeader86.352.0
Human preference scoreBenchLeader68.561.3
Knowledge scoreBenchLeader71.655.3
Long context scoreBenchLeader64.762.3
Maths scoreBenchLeader70.670.6
Multimodal scoreBenchLeader67.956.9
Reasoning scoreBenchLeader74.151.9
Instruction following scoreBenchLeader–65.1
Blended price $/MBenchLeader$10.00$3.50
Output speedBenchLeader51 tok/s166 tok/s
Time to first answerBenchLeader68.7 s6.9 s
Context windowBenchLeader1,000,000 tokens200,000 tokens
Reasoning
GPQA DiamondEpoch AI Benchmarking Hub93.9%81.8%†
Humanity's Last ExamScale AI / CAIS–20.3%†
SimpleBenchSimpleBench80.6%†53.1%†
Arena Hard PromptsArena (LMArena)15151442
LiveBench 1.0 ReasoningLiveBench91.2%–
GPQA Diamond (AA)Artificial Analysis93.2%82.7%
Humanity's Last Exam (AA)Artificial Analysis54.9%20.1%
GPQA Diamond (Vals)Vals AI93.4%†84.1%†
Kagi LLM BenchmarkKagi LLM Benchmark–67.6%
GPQA Diamond (HELM)HELM Capabilities–75.3%
ARC-AGI-1ARC Prize97.5%60.8%†
ARC-AGI-2ARC Prize90.4%6.5%†
ARC-AGI-3ARC Prize30.2%†–
CritPtArtificial Analysis29.1%1.1%
MysteryMechanismVals AI37.4%†–
SciPredictScale AI SEAL–17.9%
Chess PuzzlesEpoch AI Benchmarking Hub43.0%36.0%†
Mystery Game PuzzlesEpoch AI Benchmarking Hub59.0%29.0%†
CL-benchEpoch AI Benchmarking Hub–17.8%†
LMCAEpoch AI Benchmarking Hub63.3%39.7%†
DTBenchEpoch AI Benchmarking Hub97.6%84.8%†
ForecastBenchForecastBench66.5%64.6%
BTF-3Epoch AI Benchmarking Hub12.0%†–
Coding
SWE-bench Verified (Epoch)Epoch AI Benchmarking Hub–62.3%†
SciCodeSciCode56.4%–
Show all 166 benchmarks
MetricPublisherClaude Opus 5 (max)o3
Coding
WeirdMLWeirdML91.8%52.4%†
FrontierCodeCognition53.4%–
GSO-BenchGSO-Bench–8.8%†
Arena CodingArena (LMArena)15281460
Arena WebDevArena (LMArena)1691–
LiveBench 1.0 CodingLiveBench81.5%–
LiveCodeBenchVals AI89.0%†83.9%†
IOIVals AI84.3%†–
SWE-bench (Vals)Vals AI97.0%†–
Aider PolyglotAider polyglot leaderboard–81.3%
SWE-bench Verified (bash only)SWE-bench–58.4%
SWE-bench Verified (any scaffold)SWE-bench–58.4%
Code MigrationVals AI57.5%†–
ProgramBenchVals AI3.0%†–
Vibe Code Bench 1-100Vals AI28.5%†–
Vibe Code Bench v1.1Vals AI88.4%†–
SWE Atlas: Codebase QnAScale AI SEAL63.2%†–
SWE Atlas: Test WritingScale AI SEAL62.2%†–
CadEvalEpoch AI Benchmarking Hub–74.0%†
DeepSWE v1.1Epoch AI Benchmarking Hub73.7%–
CursorBenchCursor46.6%–
ALE-BenchEpoch AI Benchmarking Hub2164.6†933.5†
FrontierSWEEpoch AI Benchmarking Hub52.0%–
Terminal-Bench 4.0 (AA)Artificial Analysis49.0%–
Terminal-Bench 2.1 (AA)Artificial Analysis89.1%–
Startrise FrontendStartrise82.3%†–
Agents & tools
Terminal-BenchTerminal-Bench51.8%–
OSWorld-Verified 2.0OSWorld31.4%–
GDPvalOpenAI–30.8%†
APEX-AgentsMercor65.8%–
Arena AgentArena (LMArena)8.2–
LiveBench Agentic CodingLiveBench65.2%–
Terminal-Bench HardArtificial Analysis–37.1%
τ²-Bench Telecom (AA)Artificial Analysis–80.7%
Terminal-Bench 2.1 (Vals)Vals AI84.6%†–
MCP AtlasScale AI SEAL85.8%†–
HiL-BenchScale AI SEAL57.0%†–
BFCL OverallBerkeley Function Calling Leaderboard–63.0%
τ³-bench Bankingτ²-bench48.7%–
τ³-Banking (AA)Artificial Analysis42.1%–
Analyst Agent (AA)Artificial Analysis53.8%–
CUA-benchVals AI9.0%–
CyberBenchVals AI65.4%†–
Finance Agent v2Vals AI58.6%†–
Harvey's Legal Agent Benchmark v1Vals AI6.7%†–
Legal Research BenchVals AI55.3%†–
SkillsBenchVals AI60.4%†–
SREBenchVals AI12.2%†–
Tax Agent BenchVals AI45.5%†–
Terminal-Bench 4.0 (Vals)Vals AI53.5%†–
Terminal-Bench ScienceVals AI27.1%†–
Time Horizon Index: KSPVals AI18.8%†–
Arena SearchArena (LMArena)1214†–
BALROGBALROG63.4%–
DeepResearch BenchDeepResearch Bench–45.2%†
PostTrainBenchEpoch AI Benchmarking Hub35.0%†–
METR Time HorizonsMETR–63.6%
GBAEvalEpoch AI Benchmarking Hub79.6%†–
Vending-Bench 2Andon Labs11181.9†–
AutomationBenchArtificial Analysis56.6%–
GDP.pdf (AA)Artificial Analysis21.6%–
EnterpriseOps-GymArtificial Analysis47.5%–
AA-Briefcase v1.1Artificial Analysis1657–
Terminal-Bench-Science 0.1 (AA)Artificial Analysis28.6%–
DeepSec-Bench (AA)Artificial Analysis25.9%–
BrowseComp-PlusBrowseComp-Plus–50.5%
FACTS SearchGoogle85.3%†–
Maths
FrontierMath Tiers 1–3Epoch AI Benchmarking Hub85.6%33.3%†
FrontierMath Tier 4Epoch AI Benchmarking Hub73.2%–
OTIS Mock AIMEEpoch AI Benchmarking Hub98.9%83.9%†
MATH Level 5Epoch AI Benchmarking Hub–97.8%†
ProofBenchVals AI99.0%–
LiveBench MathematicsLiveBench95.7%–
AIME (Vals)Vals AI–85.3%†
MGSMVals AI–91.8%†
IMO 2025MathArena–16.7%†
Omni-MATH (HELM)HELM Capabilities–71.4%
MATH 500Vals AI–94.6%†
Arena MathsArena (LMArena)15171448
Knowledge
SimpleQA VerifiedEpoch AI Benchmarking Hub59.9%49.4%†
LiveBench 1.0 Data AnalysisLiveBench74.5%–
AA-OmniscienceArtificial Analysis37.1-15.6
MMLU-ProVals AI91.6%†85.6%†
LegalBenchVals AI87.0%†83.8%†
CorpFinVals AI73.2%†59.7%†
TaxEvalVals AI75.1%†74.6%†
MedQAVals AI–96.1%†
PRBench FinanceScale AI SEAL–47.7%
PRBench LegalScale AI SEAL–48.6%
MultiNRCScale AI SEAL–45.5%†
MMLU-Pro (HELM)HELM Capabilities–85.9%
BioMysteryBenchVals AI79.3%†–
Excel Modeling BenchmarkVals AI73.6%†–
MedCodeVals AI63.6%†47.3%†
MedScribeVals AI91.0%†76.7%†
Public Benefits Bench v1.1Vals AI76.9%†–
DrugDiscoveryBenchScale AI SEAL53.7%–
EBR-benchEpoch AI Benchmarking Hub45.7%–
GDP.pdfEpoch AI Benchmarking Hub24.0%–
AA-Omniscience: accuracyArtificial Analysis60.9%38.6%
AA-Omniscience: non-hallucinationArtificial Analysis39.2%11.9%
Instruction following
LiveBench LanguageLiveBench88.7%–
LiveBench Instruction FollowingLiveBench63.8%–
IFBenchArtificial Analysis–71.4%
MultiChallengeScale AI SEAL–56.6%†
TutorBenchScale AI SEAL–52.1%†
IFEval (HELM)HELM Capabilities–86.9%
WildBench (HELM)HELM Capabilities–86.1%
Arena Instruction FollowingArena (LMArena)14951403
Human preference
Arena TextArena (LMArena)14911432
EQ-Bench 4EQ-Bench1360†–
Arena Creative WritingArena (LMArena)14741383
Arena Multi-turnArena (LMArena)14871420
Arena Longer QueriesArena (LMArena)15031411
Lech Mazur Creative WritingLech Mazur–83.9%†
Multimodal
Arena VisionArena (LMArena)1320†1214
MMMU-ProArtificial Analysis84.7%70.1%
VISTAScale AI SEAL–50.1%†
MMMU (validation)MMMU–82.9%
MMMU-Pro (official)MMMU–76.4%
MMMU-Pro (Vals)Vals AI89.9%†80.4%†
MortgageTaxVals AI72.1%†65.7%†
SAGEVals AI49.4%†41.8%†
VTBScale AI SEAL–13.7%
GeoBenchGeoBench–60.0%†
VPCTEpoch AI Benchmarking Hub–52.0%†
Blueprint-Bench 2Epoch AI Benchmarking Hub30.4%†–
Long context
Fiction.LiveBench 120kFiction.live–100.0%†
AA-LCRArtificial Analysis79.3%74.7%
Arena DocumentArena (LMArena)1490†–
MLCRArtificial Analysis11.1%–
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub163.9†146.7
LiveBench 1.0LiveBench80.1%–
AA Intelligence Index v4.3.2Artificial Analysis50.820.2
HELM Capabilities meanHELM Capabilities–81.1%
Vals RSI IndexVals AI33.0%–
Vals Science IndexVals AI47.9%†–
Vals Multimodal IndexVals AI73.9%†–
Safety & honesty
FORTRESSScale AI SEAL16.8%†16.0%
MASKScale AI SEAL–84.5%†
PropensityBenchScale AI SEAL–10.5%

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 5 (max): 1 new result (Vals RSI Index).
  • Claude Opus 5 (max): 1 score revision.
  • Claude Opus 5 (max): 1 new result (ForecastBench); 1 score revision.
  • o3: 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…).
  • Claude Opus 5 (max): 1 score revision.
  • Claude Opus 5 (max): 2 score revisions.
  • o3: 1 new result (PRBench Legal).
  • Claude Opus 5 (max): 2 new results (Terminal-Bench-Science 0.1 (AA), DeepSec-Bench (AA)).

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

Claude Opus 5 vs o3: questions

Is Claude Opus 5 better than o3?
Claude Opus 5 (max) leads on quality: 68.8 vs 59.5. The BenchLeader Index combines every independent quality benchmark; Claude Opus 5 (max) is ahead overall as of 2026-10-26, but check the category scores for your use.
Which is cheaper, Claude Opus 5 or o3, and will that change?
o3 is cheaper today: $3.50 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 5 better than o3 for coding?
Claude Opus 5 scores higher in coding (70 vs 62 on the category index, where 50 is average). On Arena Coding, the top shared coding benchmark, it is 1528 for Claude Opus 5 against 1460 for o3.
Is Claude Opus 5 better than o3 for agents and tool use?
o3 scores higher in agentic tasks (70 vs 64 on the category index, where 50 is average).
Which is better for long documents?
Claude Opus 5 scores higher on long-context benchmarks (65 vs 62). Claude Opus 5 accepts more context: 1,000,000 against 200,000 tokens for o3.
Which has lower latency?
o3 answers first: 6.9 s against 68.7 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 o3?
o3 streams faster: 166 against 51 output tokens per second.
Do independent benchmarks agree on Claude Opus 5 vs o3?
Across the 16 benchmarks both have, Claude Opus 5 leads 16, o3 leads 0. 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.