BenchLeader
Menu
Best LLMsChoose a modelLeaderboardsModelsBenchmarksProvidersReleases9

GPT-5.4 Pro vs GPT-5.5

Full profiles: GPT-5.4 Pro (OpenAI) and GPT-5.5 (OpenAI): every benchmark, price and rank.

Updated · Data from 11 sources

As of 25 Oct 2026, GPT-5.4 Pro and GPT-5.5 (xhigh) are neck and neck on quality (63.3 vs 66.1 on the BenchLeader Index, inside the ±5.0 margin of error), and GPT-5.5 (xhigh) costs 6.0× less, which makes it the better buy for most workloads. Pick GPT-5.4 Pro for knowledge (74 vs 65); pick GPT-5.5 for instruction following (74 vs 64).

GPT-5.4 Pro benchmarks OpenAI 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
ReasoningGPT-5.576 vs 73 category scoreMedium
KnowledgeGPT-5.4 Pro74 vs 65 category scoreHigh
Images and other mediaGPT-5.4 Pro70 vs 63 category scoreMedium
Following instructionsGPT-5.574 vs 64 category scoreHigh
Lowest costGPT-5.5$11.25 vs $67.50 per 1M blended; agentic coding session $0.420 vs $2.52–
Fastest first answerGPT-5.4 Pro6.1 s (GPT-5.4 Pro) 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

  • GPT-5.4 Pro and GPT-5.5 (xhigh) are neck and neck on quality (63.3 vs 66.1, within the ±5.0 margin).
  • GPT-5.4 Pro is stronger in knowledge, multimodal.
  • GPT-5.5 (xhigh) is stronger in instruction following, reasoning, agents & tools, coding, composite, long context, maths.
  • GPT-5.5 (xhigh) is 6.0× cheaper ($11.25 vs $67.50 per 1M blended).
  • GPT-5.5 (xhigh) streams 12.3× faster (86 vs 7 tokens per second).

At a glance

SpecGPT-5.4 ProGPT-5.5
DeveloperOpenAIOpenAI
Released4 Mar 202623 Apr 2026
Context window (tokens)1,050,0001,050,000
Compared atGPT-5.4 ProGPT-5.5 (xhigh)
Open weightsNoNo
Price in / out / cached input per 1M$30.00 / $180.00 / –$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-5.4 ProGPT-5.5 (xhigh)Leader
Reasoning72.575.5GPT-5.5
Coding–61.1not comparable
Agents & tools–62.3not comparable
Maths–69.7not comparable
Knowledge73.665.3GPT-5.4 Pro
Instruction following63.873.9GPT-5.5
Multimodal69.862.9GPT-5.4 Pro
Long context–67.3not comparable
Composite–72.4not comparable

Where sources disagree

Across the 1 benchmarks both have, GPT-5.4 Pro leads 0, GPT-5.5 leads 1.

Overall indexGPT-5.4 ProGPT-5.5 (xhigh)Leader
BenchLeader Index63.366.1neck 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; suite cost is what Artificial Analysis paid to run its whole Intelligence Index.

MeasureGPT-5.4 ProGPT-5.5 (xhigh)
Blended price per 1M tokens$67.50$11.25
Input / output per 1M$30.00 / $180.00$5.00 / $30.00
Cached input per 1M–$0.500
Price per task: chat reply$0.066$0.011
Price per task: summarize a 30-page report$0.468$0.078
Price per task: code edit$0.450$0.075
Price per task: agentic coding session$2.52$0.420
Price per task: structured extraction$0.096$0.016
Cost to run a full benchmark suite–$2.63
Output speed7 tok/s86 tok/s
Time to first answer6.1 s39.1 s
Time to a complete answer–45.0 s

Each model is compared at its best-ranked reasoning-effort setting. GPT-5.4 Pro scores 63.3 and takes 6 s to its first answer token; GPT-5.4 Pro (xhigh) scores 61.9 in 6 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

137 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-5.4 ProGPT-5.5 (xhigh)
BenchLeader IndexBenchLeader63.366.1
Instruction following scoreBenchLeader63.873.9
Knowledge scoreBenchLeader73.665.3
Multimodal scoreBenchLeader69.862.9
Reasoning scoreBenchLeader72.575.5
Agents & tools scoreBenchLeader–62.3
Coding scoreBenchLeader–61.1
Composite scoreBenchLeader–72.4
Long context scoreBenchLeader–67.3
Maths scoreBenchLeader–69.7
Blended price $/MBenchLeader$67.50$11.25
Output speedBenchLeader7 tok/s86 tok/s
Time to first answerBenchLeader6.1 s39.1 s
Context windowBenchLeader1,050,000 tokens1,050,000 tokens
Reasoning
GPQA DiamondEpoch AI Benchmarking Hub94.6%†90.7%†
Humanity's Last ExamScale AI / CAIS44.3%–
SimpleBenchSimpleBench74.1%69.0%†
Arena Hard PromptsArena (LMArena)–1499†
LiveBench 2.0 ReasoningLiveBench–89.7%†
LiveBench 1.0 ReasoningLiveBench–89.7%
GPQA Diamond (AA)Artificial Analysis–93.5%
Humanity's Last Exam (AA)Artificial Analysis–45.8%
GPQA Diamond (Vals)Vals AI–93.2%
Kagi LLM BenchmarkKagi LLM Benchmark–88.8%†
ARC-AGI-1ARC Prize94.5%†95.0%
ARC-AGI-2ARC Prize83.3%†85.0%
ARC-AGI-3ARC Prize–0.4%†
CritPtArtificial Analysis30.0%†27.1%
Chess PuzzlesEpoch AI Benchmarking Hub59.0%†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–64.4%†
BTF-3Epoch AI Benchmarking Hub–14.3%†
Coding
SciCodeSciCode–55.8%
WeirdMLWeirdML57.4%†84.9%
FrontierCodeCognition–43.0%†
GSO-BenchGSO-Bench–40.2%
Show all 137 benchmarks
MetricPublisherGPT-5.4 ProGPT-5.5 (xhigh)
Coding
Arena CodingArena (LMArena)–1511†
Arena WebDevArena (LMArena)–1513
LiveBench 2.0 CodingLiveBench–85.1%†
LiveBench 1.0 CodingLiveBench–82.2%
LiveCodeBenchVals AI–85.3%
SWE-bench (Vals)Vals AI–82.6%
Code MigrationVals AI–45.2%
ProgramBenchVals AI–0.5%
Vibe Code Bench v1.1Vals AI–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 Hub–1943.0
Terminal-Bench 4.0 (AA)Artificial Analysis–14.7%
Terminal-Bench 2.1 (AA)Artificial Analysis–84.3%
Agents & tools
Terminal-BenchTerminal-Bench–84.7%†
OSWorld-Verified 2.0OSWorld–13.0%
Remote Labor IndexScale AI / CAIS–6.3%†
APEX-AgentsMercor–55.1%†
LiveBench Long AgenticLiveBench–44.6%†
LiveBench BrowserLiveBench–48.0%†
LiveBench Agentic CodingLiveBench–54.0%
Terminal-Bench HardArtificial Analysis–60.6%
τ²-Bench Telecom (AA)Artificial Analysis–93.9%
Terminal-Bench 2.1 (Vals)Vals AI–76.4%†
MCP AtlasScale AI SEAL–75.3%
HiL-BenchScale AI SEAL–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 AI–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%†
GBAEvalEpoch AI Benchmarking Hub–53.2%†
Vending-Bench 2Andon Labs–7523.8†
AutomationBenchArtificial Analysis–47.3%
GDP.pdf (AA)Artificial Analysis–21.2%
EnterpriseOps-GymArtificial Analysis–46.6%
AA-Briefcase v1.1Artificial Analysis–1136
Maths
FrontierMath Tiers 1–3Epoch AI Benchmarking Hub82.5%†85.3%
FrontierMath Tier 4Epoch AI Benchmarking Hub58.5%†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 ApexMathArena69.8%†80.2%
Arena MathsArena (LMArena)–1501†
Knowledge
SimpleQA VerifiedEpoch AI Benchmarking Hub46.3%†63.0%
LiveBench 2.0 Data AnalysisLiveBench–81.6%†
LiveBench ResearchLiveBench–61.6%†
LiveBench 1.0 Data AnalysisLiveBench–81.6%
AA-OmniscienceArtificial Analysis–20.5
MMLU-ProVals AI–88.1%
LegalBenchVals AI–86.5%
CorpFinVals AI–68.4%
TaxEvalVals AI–75.0%
MultiNRCScale AI SEAL62.3%–
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%
AA-Omniscience: accuracyArtificial Analysis–58.0%
AA-Omniscience: non-hallucinationArtificial Analysis–11.0%
Instruction following
LiveBench LanguageLiveBench–87.4%
LiveBench Instruction FollowingLiveBench–70.7%
IFBenchArtificial Analysis–75.8%
MultiChallengeScale AI SEAL69.2%–
TutorBenchScale AI SEAL56.6%–
Arena Instruction FollowingArena (LMArena)–1475†
Human preference
Arena TextArena (LMArena)–1477†
EQ-Bench 4EQ-Bench–1294†
Arena Creative WritingArena (LMArena)–1448†
Arena Multi-turnArena (LMArena)–1484†
Arena Longer QueriesArena (LMArena)–1487†
Multimodal
Arena VisionArena (LMArena)–1297†
LiveBench DesignLiveBench–70.3%†
MMMU-ProArtificial Analysis–79.9%
VISTAScale AI SEAL53.9%–
MMMU-Pro (Vals)Vals AI–88.3%
MortgageTaxVals AI–68.8%
SAGEVals AI–51.5%
Blueprint-Bench 2Epoch AI Benchmarking Hub–36.2%†
Long context
AA-LCRArtificial Analysis–84.3%
Arena DocumentArena (LMArena)–1483†
MLCRArtificial Analysis–28.9%
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub159.6159.7†
LiveBench 2.0LiveBench–68.7%†
LiveBench 1.0LiveBench–80.2%
AA Intelligence Index v4.3.2Artificial Analysis–38.4
Vals RSI IndexVals AI–16.5%
Vals Multimodal IndexVals AI–68.1%
Safety & honesty
FORTRESSScale AI SEAL14.8%16.3%
MASKScale AI SEAL91.7%–

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).
  • GPT-5.5 (xhigh): 2 new results (Harvey's Legal Agent Benchmark v1, Vals RSI Index).
  • GPT-5.4 Pro: 1 score revision.
  • GPT-5.5 (xhigh): 1 score revision.
  • GPT-5.5 (xhigh): 6 new results (USAMO 2026, AutomationBench, GDP.pdf (AA)…).
  • GPT-5.5 (xhigh): BenchLeader Index 66.2 → 65.7.
  • GPT-5.5 (xhigh): 1 score revision.
  • GPT-5.5 (xhigh): 1 new result (Chess Puzzles); 1 score revision.
  • GPT-5.5 (xhigh): 1 new result (SWE Atlas: Codebase QnA).
  • GPT-5.5 (xhigh): 1 new result (DrugDiscoveryBench).

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

GPT-5.4 Pro vs GPT-5.5: questions

Is GPT-5.4 Pro better than GPT-5.5?
GPT-5.4 Pro and GPT-5.5 (xhigh) are neck and neck on quality (63.3 vs 66.1, within the ±5.0 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, GPT-5.4 Pro or GPT-5.5, and will that change?
GPT-5.5 is cheaper today: $11.25 against $67.50 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
Which is better for long documents?
Both accept 1,050,000 tokens of context.
Which has lower latency?
GPT-5.4 Pro answers first: 6.1 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, GPT-5.4 Pro or GPT-5.5?
GPT-5.5 streams faster: 86 against 7 output tokens per second.
Do independent benchmarks agree on GPT-5.4 Pro vs GPT-5.5?
Across the 1 benchmarks both have, GPT-5.4 Pro leads 0, GPT-5.5 leads 1.
When was GPT-5.5 released?
GPT-5.5 was released on 23 Apr 2026.