BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases9

GPT-5.5 vs o3

Full profiles: GPT-5.5 (OpenAI) and o3 (OpenAI): every benchmark, price and rank.

Updated · Data from 18 sources

As of 29 Oct 2026, GPT-5.5 is ahead of o3 on quality (64.4 vs 61.0 on the BenchLeader Index), but o3 is 3.2× cheaper at $3.50 vs $11.25 per 1M tokens blended. Choose GPT-5.5 for instruction following (74 vs 65), reasoning (69 vs 61), and knowledge (64 vs 56); choose o3 for fast first answers (5.9 s vs 41.3 s) and agents & tools (70 vs 62), where it also streams ~1.5× faster.

Key facts (as of 29 Oct 2026)

  • GPT-5.5 scores 64.4 on the BenchLeader Index and o3 scores 61.0 (BenchLeader, 29 Oct 2026).
  • GPT-5.5 costs $11.25 per 1M tokens blended (list price, 29 Oct 2026).
  • o3 costs $3.50 per 1M tokens blended (list price, 29 Oct 2026).
  • o3 streams ~1.5× faster than GPT-5.5 (Artificial Analysis, 7-day median, 29 Oct 2026).
  • Of the 10 benchmarks both models have, GPT-5.5 leads 10 (BenchLeader, 29 Oct 2026).
GPT-5.5 benchmarks xhigh OpenAI 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
CodingEither61 vs 62 category scoreEither
Agents & tool useo370 vs 62 category score; Terminal-Bench Hard 37.1% vs 60.6%High
MathsEither70 vs 71 category scoreEither
ReasoningGPT-5.569 vs 61 category score; GPQA Diamond (AA) 93.5% vs 82.7%High
KnowledgeGPT-5.564 vs 56 category scoreHigh
Long documentsGPT-5.567 vs 62 category score; AA-LCR v1.1 84.3% vs 74.7%Medium
Images and other mediaGPT-5.563 vs 57 category score; MMMU-Pro 79.9% vs 70.1%Medium
Following instructionsGPT-5.574 vs 65 category score; IFBench 75.8% vs 71.4%High
Chat quality (human preference)–Only o3 has an Arena rating (61 category score)Not comparable
Lowest costo3$3.50 vs $11.25 per 1M blended; agentic coding session $0.152 vs $0.420–
Fastest first answero35.9 s (o3) vs 41.3 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-5.5 (xhigh) leads on quality: 64.4 vs 61.0.
  • GPT-5.5 (xhigh) is stronger in instruction following, knowledge, long context, multimodal, reasoning.
  • o3 is stronger in agents & tools, coding, maths, human preference.
  • o3 is 3.2× cheaper ($3.50 vs $11.25 per 1M blended).
  • o3 streams 1.8× faster (158 vs 87 tokens per second).

At a glance

SpecGPT-5.5o3
DeveloperOpenAIOpenAI
Released23 Apr 202616 Apr 2025
Context window (tokens)1,050,000200,000
Compared atGPT-5.5 (xhigh)o3
Open weightsNoNo
Price in / out / cached input per 1M$5.00 / $30.00 / $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.

CategoryGPT-5.5 (xhigh)o3Leader
Reasoning68.960.9GPT-5.5
Coding60.961.8level
Agents & tools62.370.3o3
Maths69.670.5level
Knowledge63.756.1GPT-5.5
Instruction following73.965.1GPT-5.5
Human preference–61.1not comparable
Multimodal62.656.7GPT-5.5
Long context67.162.1GPT-5.5

Where sources disagree

Across the 10 benchmarks both have, GPT-5.5 leads 10, o3 leads 0. Every overall index that rates both puts GPT-5.5 ahead.

Overall indexGPT-5.5 (xhigh)o3Leader
BenchLeader Index64.461.0GPT-5.5
Arena Text1482†1432GPT-5.5

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.

MeasureGPT-5.5 (xhigh)o3
Blended price per 1M tokens$11.25$3.50
Input / output per 1M$5.00 / $30.00$2.00 / $8.00
Cached input per 1M$0.500$0.500
Price per task: chat reply$0.011$0.0032
Price per task: summarize a 30-page report$0.078$0.029
Price per task: code edit$0.075$0.024
Price per task: agentic coding session$0.420$0.152
Price per task: structured extraction$0.016$0.0056
Output speed87 tok/s158 tok/s
Time to first answer41.3 s5.9 s
Time to a complete answer47.1 s9.1 s

Each model is compared at its best-ranked reasoning-effort setting. GPT-5.5 (xhigh) scores 64.4 and takes 41 s to its first answer token; GPT-5.5 (high) scores 63.7 in 16 s, GPT-5.5 scores 61.7 in 3 s. o3 scores 61.0 and takes 6 s to its first answer token; o3 (high) scores 53.8 in 6 s, o3 (medium) scores 53.3 in 6 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

162 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-5.5 (xhigh)o3
BenchLeader IndexBenchLeader64.461.0
Agents & tools scoreBenchLeader62.370.3
Coding scoreBenchLeader60.961.8
Instruction following scoreBenchLeader73.965.1
Knowledge scoreBenchLeader63.756.1
Long context scoreBenchLeader67.162.1
Maths scoreBenchLeader69.670.5
Multimodal scoreBenchLeader62.656.7
Reasoning scoreBenchLeader68.960.9
Human preference scoreBenchLeader–61.1
Blended price $/MBenchLeader$11.25$3.50
Output speedBenchLeader87 tok/s158 tok/s
Time to first answerBenchLeader41.3 s5.9 s
Context windowBenchLeader1,050,000 tokens200,000 tokens
Reasoning
GPQA DiamondEpoch AI Benchmarking Hub90.7%†81.8%†
Humanity's Last ExamScale AI / CAIS–20.3%†
SimpleBenchSimpleBench69.0%†53.1%†
Arena Hard PromptsArena (LMArena)1499†1441
LiveBench 1.0 ReasoningLiveBench89.7%–
GPQA Diamond (AA)Artificial Analysis93.5%82.7%
Humanity's Last Exam (AA)Artificial Analysis45.8%20.1%
GPQA Diamond (Vals)Vals AI93.2%84.1%†
Kagi LLM BenchmarkKagi LLM Benchmark88.8%†67.6%
GPQA Diamond (HELM)HELM Capabilities–75.3%
ARC-AGI-1ARC Prize95.0%60.8%†
ARC-AGI-2ARC Prize85.0%6.5%†
ARC-AGI-3ARC Prize0.4%†–
CritPtArtificial Analysis27.1%1.1%
SciPredictScale AI SEAL–17.9%
Chess PuzzlesEpoch AI Benchmarking Hub63.0%36.0%†
Mystery Game PuzzlesEpoch AI Benchmarking Hub56.0%29.0%†
CL-benchEpoch AI Benchmarking Hub–17.8%†
CL-bench LifeEpoch AI Benchmarking Hub22.2%†–
LMCAEpoch AI Benchmarking Hub54.3%39.7%†
DTBenchEpoch AI Benchmarking Hub96.0%84.8%†
ForecastBenchForecastBench64.5%†64.6%
BTF-3Epoch AI Benchmarking Hub14.3%†–
Coding
SWE-bench Verified (Epoch)Epoch AI Benchmarking Hub–62.3%†
SciCodeSciCode55.8%–
Show all 162 benchmarks
MetricPublisherGPT-5.5 (xhigh)o3
Coding
WeirdMLWeirdML84.9%52.4%†
FrontierCodeCognition43.0%†–
GSO-BenchGSO-Bench40.2%8.8%†
Arena CodingArena (LMArena)1510†1460
Arena WebDevArena (LMArena)1512–
LiveBench 1.0 CodingLiveBench82.2%–
LiveCodeBenchVals AI85.3%83.9%†
SWE-bench (Vals)Vals AI82.6%–
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 AI45.2%–
ProgramBenchVals AI0.5%–
Vibe Code Bench v1.1Vals AI69.8%–
SWE Atlas: Codebase QnAScale AI SEAL45.4%–
SWE Atlas: RefactoringScale AI SEAL44.8%–
SWE Atlas: Test WritingScale AI SEAL42.6%–
MirrorCodeEpoch AI Benchmarking Hub10.0%†–
CadEvalEpoch AI Benchmarking Hub–74.0%†
Surface Evolver BenchEpoch AI Benchmarking Hub88.1%†–
DeepSWE v1.1Epoch AI Benchmarking Hub67.0%–
ALE-BenchEpoch AI Benchmarking Hub1943.0933.5†
Terminal-Bench 4.0 (AA)Artificial Analysis14.7%–
Terminal-Bench 2.1 (AA)Artificial Analysis84.3%–
LiveCodeBench (AA)Artificial Analysis–80.8%
Agents & tools
Terminal-BenchTerminal-Bench84.7%†–
OSWorld-Verified 2.0OSWorld13.0%–
GDPvalOpenAI–30.8%†
Remote Labor IndexScale AI / CAIS6.3%†–
APEX-AgentsMercor55.1%†–
LiveBench Agentic CodingLiveBench54.0%–
Terminal-Bench HardArtificial Analysis60.6%37.1%
τ²-Bench Telecom (AA)Artificial Analysis93.9%80.7%
Terminal-Bench 2.1 (Vals)Vals AI76.4%†–
MCP AtlasScale AI SEAL75.3%–
HiL-BenchScale AI SEAL39.7%†–
BFCL OverallBerkeley Function Calling Leaderboard–63.0%
τ³-bench Bankingτ²-bench44.6%–
τ³-Banking (AA)Artificial Analysis39.0%–
ITBench SRE (AA)Artificial Analysis45.8%–
Analyst Agent (AA)Artificial Analysis50.0%–
APEX-Agents (AA)Artificial Analysis37.7%–
Finance Agent v2Vals AI51.8%–
Harvey's Legal Agent Benchmark v1Vals AI3.8%–
Legal Research BenchVals AI40.4%–
SkillsBenchVals AI62.2%–
SREBenchVals AI3.8%–
Tax Agent BenchVals AI21.8%–
Terminal-Bench 2.0 (Vals)Vals AI73.2%†–
Time Horizon Index: KSPVals AI8.3%–
DeepResearch BenchDeepResearch Bench54.0%†45.2%†
PostTrainBenchEpoch AI Benchmarking Hub27.2%–
ExploitBenchEpoch AI Benchmarking Hub47.4%†–
METR Time HorizonsMETR–63.6%
GBAEvalEpoch AI Benchmarking Hub53.2%†–
Vending-Bench 2Andon Labs7523.8†–
AutomationBenchArtificial Analysis47.3%–
GDP.pdf (AA)Artificial Analysis21.2%–
EnterpriseOps-GymArtificial Analysis46.6%–
AA-Briefcase v1.1Artificial Analysis1136–
BrowseComp-PlusBrowseComp-Plus–50.5%
Maths
FrontierMath Tiers 1–3Epoch AI Benchmarking Hub85.3%33.3%†
FrontierMath Tier 4Epoch AI Benchmarking Hub72.5%–
OTIS Mock AIMEEpoch AI Benchmarking Hub84.4%†83.9%†
MATH Level 5Epoch AI Benchmarking Hub–97.8%†
ProofBenchVals AI50.0%–
LiveBench MathematicsLiveBench95.9%–
AIME (Vals)Vals AI–85.3%†
MGSMVals AI–91.8%†
AIME 2026MathArena100.0%–
HMMT February 2026MathArena98.5%–
USAMO 2026MathArena98.2%–
IMO 2025MathArena–16.7%†
MathArena ApexMathArena80.2%–
Omni-MATH (HELM)HELM Capabilities–71.4%
MATH 500Vals AI–94.6%†
Arena MathsArena (LMArena)1500†1448
AIME 2025 (AA)Artificial Analysis–88.3%
Knowledge
SimpleQA VerifiedEpoch AI Benchmarking Hub63.0%49.4%†
LiveBench 1.0 Data AnalysisLiveBench81.6%–
MMLU-ProVals AI88.1%85.6%†
LegalBenchVals AI86.5%83.8%†
CorpFinVals AI68.4%59.7%†
TaxEvalVals AI75.0%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%
CaseLaw v2Vals AI66.2%–
Excel Modeling BenchmarkVals AI64.5%–
MedCodeVals AI49.1%47.3%†
MedScribeVals AI86.9%76.7%†
Public Benefits Bench v1.1Vals AI60.9%–
DrugDiscoveryBenchScale AI SEAL51.6%–
EBR-benchEpoch AI Benchmarking Hub34.3%–
GDP.pdfEpoch AI Benchmarking Hub26.0%–
Instruction following
LiveBench LanguageLiveBench87.4%–
LiveBench Instruction FollowingLiveBench70.7%–
IFBenchArtificial Analysis75.8%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)1475†1403
Human preference
Arena TextArena (LMArena)1477†1432
EQ-Bench 4EQ-Bench1295†–
Arena Creative WritingArena (LMArena)1448†1383
Arena Multi-turnArena (LMArena)1484†1420
Arena Longer QueriesArena (LMArena)1487†1411
Lech Mazur Creative WritingLech Mazur–83.9%†
Multimodal
Arena VisionArena (LMArena)1297†1214
MMMU-ProArtificial Analysis79.9%70.1%
VISTAScale AI SEAL–50.1%†
MMMU (validation)MMMU–82.9%
MMMU-Pro (official)MMMU–76.4%
MMMU-Pro (Vals)Vals AI88.3%80.4%†
MortgageTaxVals AI68.8%65.7%†
VTBScale AI SEAL–13.7%
GeoBenchGeoBench–60.0%†
VPCTEpoch AI Benchmarking Hub–52.0%†
Blueprint-Bench 2Epoch AI Benchmarking Hub36.2%†–
Code2Video-AAArtificial Analysis1000–
Long context
Fiction.LiveBench 120kFiction.live–100.0%†
AA-LCR v1.1Artificial Analysis84.3%74.7%
Arena DocumentArena (LMArena)1483†–
MLCRArtificial Analysis28.9%–
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub159.6†146.8
LiveBench 1.0LiveBench80.2%–
AA Intelligence Index v4.3.2Artificial Analysis38.420.2
AA Progress Index v1Artificial Analysis320258.3
HELM Capabilities meanHELM Capabilities–81.1%
Vals RSI IndexVals AI16.5%–
Vals Multimodal IndexVals AI68.1%–
Safety & honesty
FORTRESSScale AI SEAL16.3%16.0%
MASKScale AI SEAL–84.5%†
PropensityBenchScale AI SEAL–10.5%

Data as of 2026-10-29. Best configuration of each model; every score links to its source on the model pages.

What changed

  • GPT-5.5 (xhigh): 1 new result (Code2Video-AA).
  • o3: 2 new results (LiveCodeBench (AA), AIME 2025 (AA)).
  • GPT-5.5 (xhigh): 1 new result (AA Progress Index v1).
  • o3: 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.
  • o3: 1 score revision.
  • GPT-5.5 (xhigh): BenchLeader Index 65.1 → 64.6.

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

GPT-5.5 vs o3: questions

Is GPT-5.5 better than o3?
GPT-5.5 (xhigh) leads on quality: 64.4 vs 61.0. The BenchLeader Index combines every independent quality benchmark; GPT-5.5 (xhigh) is ahead overall as of 29 Oct 2026, but check the category scores for your use.
GPT-5.5 vs o3: which is cheaper, and will that change?
o3 is cheaper today: $3.50 against $11.25 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
Is GPT-5.5 or o3 better for coding?
They are level in coding (61 vs 62 on the category index).
Which is faster, GPT-5.5 or o3?
o3 streams ~1.5× faster (158 vs 87 output tokens per second, Artificial Analysis 7-day median).
Is GPT-5.5 or o3 better for agents and tool use?
o3 scores higher in agentic tasks (70 vs 62 on the category index, where 50 is average). On Terminal-Bench Hard, the top shared agents & tools benchmark, it is 60.6% for GPT-5.5 against 37.1% for o3.
More questions (4)
Which is better for long documents?
GPT-5.5 scores higher on long-context benchmarks (67 vs 62). GPT-5.5 accepts more context: 1,050,000 against 200,000 tokens for o3.
Which has lower latency?
o3 answers first: 5.9 s against 41.3 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-5.5 vs o3?
Across the 10 benchmarks both have, GPT-5.5 leads 10, o3 leads 0. Every overall index that rates both puts GPT-5.5 ahead.
When was GPT-5.5 released?
GPT-5.5 was released on 23 Apr 2026.