BenchLeader
Menu
Best LLMsChoose a modelLeaderboardsModelsBenchmarksProvidersReleases12

GPT-5.4 Pro vs GPT-6 Astra

Full profiles: GPT-5.4 Pro (OpenAI) and GPT-6 Astra (OpenAI): every benchmark, price and rank.

Updated · Data from 10 sources

As of 24 Oct 2026, GPT-6 Astra (max) is the better pick for most workloads: it leads on quality (70.4 vs 63.3 on the BenchLeader Index) and costs 3.4× less. Pick GPT-5.4 Pro for multimodal (70 vs 66); pick GPT-6 Astra for knowledge (80 vs 74).

GPT-5.4 Pro benchmarks OpenAI GPT-6 Astra benchmarks max 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-6 Astra76 vs 73 category scoreMedium
KnowledgeGPT-6 Astra80 vs 74 category scoreMedium
Images and other mediaGPT-5.4 Pro70 vs 66 category scoreMedium
Chat quality (human preference)–Only GPT-6 Astra has an Arena rating (67 category score)Not comparable
Lowest costGPT-6 Astra$20.00 vs $67.50 per 1M blended; agentic coding session $0.800 vs $2.52–
Fastest first answerGPT-5.4 Pro6.1 s (GPT-5.4 Pro) vs 247.0 s (GPT-6 Astra (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

  • GPT-6 Astra (max) leads on quality: 70.4 vs 63.3.
  • GPT-5.4 Pro is stronger in instruction following, multimodal.
  • GPT-6 Astra (max) is stronger in knowledge, reasoning, agents & tools, coding, composite, human preference, long context, maths.
  • GPT-6 Astra (max) is 3.4× cheaper ($20.00 vs $67.50 per 1M blended).
  • GPT-6 Astra (max) streams 8.3× faster (58 vs 7 tokens per second).

At a glance

SpecGPT-5.4 ProGPT-6 Astra
DeveloperOpenAIOpenAI
Released4 Mar 20262 Sept 2026
Context window (tokens)1,050,0001,050,000
Compared atGPT-5.4 ProGPT-6 Astra (max)
Open weightsNoNo
Price in / out / cached input per 1M$30.00 / $180.00 / –$10.00 / $50.00 / $1.00

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-6 Astra (max)Leader
Reasoning72.576.2GPT-6 Astra
Coding–74.0not comparable
Agents & tools–64.2not comparable
Maths–71.8not comparable
Knowledge73.679.5GPT-6 Astra
Instruction following63.8–not comparable
Human preference–66.7not comparable
Multimodal69.866.0GPT-5.4 Pro
Long context–65.4not comparable
Composite–88.4not comparable

Where sources disagree

Across the 0 benchmarks both have, GPT-5.4 Pro leads 0, GPT-6 Astra leads 0. Every overall index that rates both puts GPT-6 Astra ahead.

Overall indexGPT-5.4 ProGPT-6 Astra (max)Leader
BenchLeader Index63.370.4GPT-6 Astra

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-6 Astra (max)
Blended price per 1M tokens$67.50$20.00
Input / output per 1M$30.00 / $180.00$10.00 / $50.00
Cached input per 1M–$1.00
Price per task: chat reply$0.066$0.019
Price per task: summarize a 30-page report$0.468$0.150
Price per task: code edit$0.450$0.135
Price per task: agentic coding session$2.52$0.800
Price per task: structured extraction$0.096$0.030
Cost to run a full benchmark suite–$3.26
Output speed7 tok/s58 tok/s
Time to first answer6.1 s247.0 s
Time to a complete answer–255.6 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-6 Astra (max) scores 70.4 and takes 247 s to its first answer token; GPT-6 Astra (high) scores 69.6 in 35 s, GPT-6 Astra (xhigh) scores 68.9 in 130 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

138 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-6 Astra (max)
BenchLeader IndexBenchLeader63.370.4
Instruction following scoreBenchLeader63.8–
Knowledge scoreBenchLeader73.679.5
Multimodal scoreBenchLeader69.866.0
Reasoning scoreBenchLeader72.576.2
Agents & tools scoreBenchLeader–64.2
Coding scoreBenchLeader–74.0
Composite scoreBenchLeader–88.4
Human preference scoreBenchLeader–66.7
Long context scoreBenchLeader–65.4
Maths scoreBenchLeader–71.8
Blended price $/MBenchLeader$67.50$20.00
Output speedBenchLeader7 tok/s58 tok/s
Time to first answerBenchLeader6.1 s247.0 s
Context windowBenchLeader1,050,000 tokens1,050,000 tokens
Reasoning
GPQA DiamondEpoch AI Benchmarking Hub94.6%†95.8%
Humanity's Last ExamScale AI / CAIS44.3%54.8%†
SimpleBenchSimpleBench74.1%–
Arena Hard PromptsArena (LMArena)–1501
LiveBench 2.0 ReasoningLiveBench–92.7%†
LiveBench 1.0 ReasoningLiveBench–92.7%
GPQA Diamond (AA)Artificial Analysis–96.1%
Humanity's Last Exam (AA)Artificial Analysis–54.7%
ARC-AGI-1ARC Prize94.5%†97.5%
ARC-AGI-2ARC Prize83.3%†95.0%
ARC-AGI-3ARC Prize–98.6%
CritPtArtificial Analysis30.0%†31.7%
MysteryMechanismVals AI–53.1%
Chess PuzzlesEpoch AI Benchmarking Hub59.0%†74.0%
Mystery Game PuzzlesEpoch AI Benchmarking Hub–84.0%
LMCAEpoch AI Benchmarking Hub–64.1%†
DTBenchEpoch AI Benchmarking Hub–97.1%†
BTF-3Epoch AI Benchmarking Hub–13.5%†
Coding
SciCodeSciCode–56.5%
WeirdMLWeirdML57.4%†93.3%
FrontierCodeCognition–53.3%
GSO-BenchGSO-Bench–79.4%†
Arena CodingArena (LMArena)–1543
Arena WebDevArena (LMArena)–1775
LiveBench 2.0 CodingLiveBench–93.7%†
Show all 138 benchmarks
MetricPublisherGPT-5.4 ProGPT-6 Astra (max)
Coding
LiveBench 1.0 CodingLiveBench–80.4%
IOIVals AI–100.0%
Code MigrationVals AI–67.7%
Code Migration v1.1Vals AI–80.6%
ProgramBenchVals AI–5.5%
SWE-bench MultimodalVals AI–43.3%
Vibe Code Bench 1-100Vals AI–27.6%
Vibe Code Bench v1.1Vals AI–89.6%
SWE Atlas: Codebase QnAScale AI SEAL–59.1%†
SWE Atlas: RefactoringScale AI SEAL–59.0%†
SWE Atlas: Test WritingScale AI SEAL–50.7%†
MirrorCodeEpoch AI Benchmarking Hub–46.7%†
DeepSWE v1.1Epoch AI Benchmarking Hub–73.2%
ALE-BenchEpoch AI Benchmarking Hub–2951.3
FrontierSWEEpoch AI Benchmarking Hub–65.5%
Terminal-Bench 4.0 (AA)Artificial Analysis–59.1%
Terminal-Bench 2.1 (AA)Artificial Analysis–88.4%
Startrise FrontendStartrise–79.9%†
FrontierCode 1.1 (AA)Artificial Analysis–47.0%
Agents & tools
Terminal-BenchTerminal-Bench–58.2%
Remote Labor IndexScale AI / CAIS–20.8%†
APEX-AgentsMercor–64.7%†
Arena AgentArena (LMArena)–14.7
LiveBench Long AgenticLiveBench–82.4%†
LiveBench BrowserLiveBench–84.0%†
LiveBench Agentic CodingLiveBench–57.3%
Terminal-Bench 2.1 (Vals)Vals AI–87.3%
HiL-BenchScale AI SEAL–65.3%†
GDPval-AA v2.1Artificial Analysis–53.7%
τ³-Banking (AA)Artificial Analysis–41.4%
ITBench SRE (AA)Artificial Analysis–48.6%
Analyst Agent (AA)Artificial Analysis–51.3%
CUA-benchVals AI–19.2%
CyberBenchVals AI–41.1%
Finance Agent v2Vals AI–53.5%
Finance Agent v2.1Vals AI–60.4%
Harvey's Legal Agent Benchmark v1.1Vals AI–7.1%
Harvey's Legal Agent Benchmark v1Vals AI–5.4%
Legal Research BenchVals AI–39.4%
Robo-benchVals AI–23.3%
SREBenchVals AI–56.9%
Tax Agent BenchVals AI–20.7%
Terminal-Bench 4.0 (Vals)Vals AI–59.6%
Terminal-Bench ScienceVals AI–62.9%
Time Horizon Index: KSPVals AI–90.5%
Arena SearchArena (LMArena)–1246
BALROGBALROG–68.3%
Vending-Bench 2Andon Labs–15514.7†
AutomationBenchArtificial Analysis–68.5%
GDP.pdf (AA)Artificial Analysis–31.0%
Harvey LAB-AA v1.1Harvey–8.6%
AA-Briefcase v1.1Artificial Analysis–1568
Terminal-Bench-Science 0.1 (AA)Artificial Analysis–63.3%
CWE-Bench (AA)Artificial Analysis–63.3%
DeepSec-Bench (AA)Artificial Analysis–36.9%
CyberGym end-to-end (AA)Artificial Analysis–0.0%
FACTS SearchGoogle–89.9%†
Maths
FrontierMath Tiers 1–3Epoch AI Benchmarking Hub82.5%†93.7%
FrontierMath Tier 4Epoch AI Benchmarking Hub58.5%†97.6%
OTIS Mock AIMEEpoch AI Benchmarking Hub–100.0%
ProofBenchVals AI–99.0%†
LiveBench MathematicsLiveBench–96.8%
MathArena ApexMathArena69.8%†–
Arena MathsArena (LMArena)–1487
Knowledge
SimpleQA VerifiedEpoch AI Benchmarking Hub46.3%†75.6%
LiveBench 2.0 Data AnalysisLiveBench–83.0%†
LiveBench ResearchLiveBench–85.2%†
LiveBench 1.0 Data AnalysisLiveBench–83.0%
AA-OmniscienceArtificial Analysis–43.4
PRBench FinanceScale AI SEAL–47.5%†
PRBench LegalScale AI SEAL–48.4%†
MultiNRCScale AI SEAL62.3%–
BioMysteryBenchVals AI–79.3%
Excel Modeling BenchmarkVals AI–71.7%
Excel Modeling Benchmark v1.1Vals AI–72.7%
MedCodeVals AI–48.5%
MedScribeVals AI–87.9%
DrugDiscoveryBenchScale AI SEAL–68.7%
EBR-benchEpoch AI Benchmarking Hub–76.2%
GDP.pdfEpoch AI Benchmarking Hub–34.2%
AA-Omniscience: accuracyArtificial Analysis–62.6%
AA-Omniscience: non-hallucinationArtificial Analysis–48.7%
Instruction following
LiveBench LanguageLiveBench–89.4%
LiveBench Instruction FollowingLiveBench–75.6%
MultiChallengeScale AI SEAL69.2%–
TutorBenchScale AI SEAL56.6%–
Arena Instruction FollowingArena (LMArena)–1473
Human preference
Arena TextArena (LMArena)–1477
Arena Creative WritingArena (LMArena)–1450
Arena Multi-turnArena (LMArena)–1485
Arena Longer QueriesArena (LMArena)–1489
Multimodal
Arena VisionArena (LMArena)–1280
LiveBench DesignLiveBench–82.0%†
MMMU-ProArtificial Analysis–86.9%
VISTAScale AI SEAL53.9%–
SAGEVals AI–46.4%
Blueprint-Bench 2Epoch AI Benchmarking Hub–49.7%†
AA-3DArtificial Analysis–1166
Long context
AA-LCRArtificial Analysis–80.7%
Arena DocumentArena (LMArena)–1473
MLCRArtificial Analysis–41.7%
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub159.6168.8†
LiveBench 2.0LiveBench–86.1%†
LiveBench 1.0LiveBench–82.2%
AA Intelligence Index v4.3.2Artificial Analysis–52.7
Vals IndexVals AI–65.0
Vals Finance IndexVals AI–55.6%
Vals RSI IndexVals AI–27.1%
Vals Science IndexVals AI–65.1%
Vals Router IndexVals AI–65.0%†
AA Cyber IndexArtificial Analysis–33.4
Safety & honesty
FORTRESSScale AI SEAL14.8%–
MASKScale AI SEAL91.7%–

Data as of 2026-10-24. Best configuration of each model; every score links to its source on the model pages.

What changed

  • GPT-6 Astra (max): 9 new results (Code Migration v1.1, Excel Modeling Benchmark v1.1, Finance Agent v2.1…).
  • GPT-5.4 Pro: 1 score revision.
  • GPT-6 Astra (max): 1 score revision.
  • GPT-6 Astra (max): 4 score revisions.
  • GPT-6 Astra (max): BenchLeader Index 69.8 → 70.6.
  • GPT-6 Astra (max): 1 score revision.
  • GPT-6 Astra (max): 1 new result (Arena Search); 3 score revisions.
  • GPT-6 Astra (max): BenchLeader Index 70.3 → 69.8.
  • GPT-6 Astra (max): 1 score revision.
  • GPT-6 Astra (max): 7 new results (FrontierCode 1.1 (AA), Terminal-Bench-Science 0.1 (AA), AA Cyber Index…); 1 score revision.

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

GPT-5.4 Pro vs GPT-6 Astra: questions

Is GPT-5.4 Pro better than GPT-6 Astra?
GPT-6 Astra (max) leads on quality: 70.4 vs 63.3. The BenchLeader Index combines every independent quality benchmark; GPT-6 Astra (max) is ahead overall as of 2026-10-24, but check the category scores for your use.
Which is cheaper, GPT-5.4 Pro or GPT-6 Astra, and will that change?
GPT-6 Astra is cheaper today: $20.00 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 247.0 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-6 Astra?
GPT-6 Astra streams faster: 58 against 7 output tokens per second.
When was GPT-6 Astra released?
GPT-6 Astra was released on 2 Sept 2026.