BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases8

GPT-5.6 Luna vs GPT-6 Astra

Full profiles: GPT-5.6 Luna (OpenAI) and GPT-6 Astra (OpenAI): every benchmark, price and rank.

Updated · Data from 11 sources

As of 29 Oct 2026, GPT-6 Astra is ahead of GPT-5.6 Luna on quality (68.2 vs 58.5 on the BenchLeader Index), but GPT-5.6 Luna is 44× cheaper at $0.450 vs $20.00 per 1M tokens blended. GPT-6 Astra leads in every category we measure both in. GPT-5.6 Luna streams 3×+ faster.

Key facts (as of 29 Oct 2026)

  • GPT-5.6 Luna scores 58.5 on the BenchLeader Index and GPT-6 Astra scores 68.2 (BenchLeader, 29 Oct 2026).
  • GPT-5.6 Luna costs $0.450 per 1M tokens blended (list price, 29 Oct 2026).
  • GPT-6 Astra costs $20.00 per 1M tokens blended (list price, 29 Oct 2026).
  • GPT-5.6 Luna streams 3×+ faster than GPT-6 Astra (Artificial Analysis, 7-day median, 29 Oct 2026).
  • Of the 2 benchmarks both models have, GPT-6 Astra leads 2 (BenchLeader, 29 Oct 2026).
GPT-5.6 Luna benchmarks xhigh OpenAI GPT-6 Astra 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
CodingGPT-6 Astra83 vs 57 category scoreHigh
Agents & tool useGPT-6 Astra70 vs 57 category scoreHigh
MathsGPT-6 Astra78 vs 66 category scoreHigh
ReasoningGPT-6 Astra86 vs 53 category score; LMCA 64.1% vs 48.5%High
Chat quality (human preference)–Only GPT-5.6 Luna has an Arena rating (64 category score)Not comparable
Lowest costGPT-5.6 Luna$0.450 vs $20.00 per 1M blended; agentic coding session $0.017 vs $0.800–
Fastest first answerGPT-6 Astra3.0 s (GPT-6 Astra) vs 38.0 s (GPT-5.6 Luna (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-6 Astra leads on quality: 68.2 vs 58.5.
  • GPT-5.6 Luna (xhigh) is stronger in human preference, long context, multimodal.
  • GPT-6 Astra is stronger in agents & tools, coding, maths, reasoning, composite, knowledge.
  • GPT-5.6 Luna (xhigh) is 44× cheaper ($0.450 vs $20.00 per 1M blended).
  • GPT-5.6 Luna (xhigh) streams 3.1× faster (124 vs 40 tokens per second).

At a glance

SpecGPT-5.6 LunaGPT-6 Astra
DeveloperOpenAIOpenAI
Released9 Jul 20262 Sept 2026
Context window (tokens)1,050,0001,050,000
Compared atGPT-5.6 Luna (xhigh)GPT-6 Astra
Open weightsNoNo
Price in / out / cached input per 1M$0.200 / $1.20 / $0.020$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-5.6 Luna (xhigh)GPT-6 AstraLeader
Reasoning52.985.6GPT-6 Astra
Coding57.083.2GPT-6 Astra
Agents & tools57.169.6GPT-6 Astra
Maths65.977.5GPT-6 Astra
Knowledge–55.7not comparable
Human preference63.8–not comparable
Multimodal60.1–not comparable
Long context65.7–not comparable
Composite–73.7not comparable

Where sources disagree

Across the 2 benchmarks both have, GPT-5.6 Luna leads 0, GPT-6 Astra leads 2. Every overall index that rates both puts GPT-6 Astra ahead.

Overall indexGPT-5.6 Luna (xhigh)GPT-6 AstraLeader
BenchLeader Index58.568.2GPT-6 Astra
Arena Text14531477†GPT-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-5.6 Luna (xhigh)GPT-6 Astra
Blended price per 1M tokens$0.450$20.00
Input / output per 1M$0.200 / $1.20$10.00 / $50.00
Cached input per 1M$0.020$1.00
Price per task: chat reply$0.0004$0.019
Price per task: summarize a 30-page report$0.0031$0.150
Price per task: code edit$0.0030$0.135
Price per task: agentic coding session$0.017$0.800
Price per task: structured extraction$0.0006$0.030
Output speed124 tok/s40 tok/s
Time to first answer38.0 s3.0 s
Time to a complete answer42.0 s–

Each model is compared at its best-ranked reasoning-effort setting. GPT-5.6 Luna (xhigh) scores 58.5 and takes 38 s to its first answer token; GPT-5.6 Luna (max) scores 57.8 in 95 s, GPT-5.6 Luna (high) scores 54.9 in 9 s. GPT-6 Astra scores 68.2 and takes 3 s to its first answer token; GPT-6 Astra (max) scores 67.7 in 232 s, GPT-6 Astra (high) scores 66.8 in 33 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

155 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-5.6 Luna (xhigh)GPT-6 Astra
BenchLeader IndexBenchLeader58.568.2
Agents & tools scoreBenchLeader57.169.6
Coding scoreBenchLeader57.083.2
Human preference scoreBenchLeader63.8–
Long context scoreBenchLeader65.7–
Maths scoreBenchLeader65.977.5
Multimodal scoreBenchLeader60.1–
Reasoning scoreBenchLeader52.985.6
Composite scoreBenchLeader–73.7
Knowledge scoreBenchLeader–55.7
Blended price $/MBenchLeader$0.450$20.00
Output speedBenchLeader124 tok/s40 tok/s
Time to first answerBenchLeader38.0 s3.0 s
Context windowBenchLeader1,050,000 tokens1,050,000 tokens
Reasoning
GPQA DiamondEpoch AI Benchmarking Hub91.6%†95.8%†
Humanity's Last ExamScale AI / CAIS–54.8%
SimpleBenchSimpleBench46.8%–
Arena Hard PromptsArena (LMArena)14751501†
LiveBench 2.0 ReasoningLiveBench–92.7%
LiveBench 1.0 ReasoningLiveBench85.6%†92.7%†
GPQA Diamond (AA)Artificial Analysis89.5%96.1%†
Humanity's Last Exam (AA)Artificial Analysis37.0%54.7%†
HLE-Diamond (AA)Artificial Analysis–63.8%†
GPQA Diamond (Vals)Vals AI91.7%†–
Kagi LLM BenchmarkKagi LLM Benchmark49.1%†–
ARC-AGI-1ARC Prize87.7%97.5%†
ARC-AGI-2ARC Prize47.6%95.0%†
ARC-AGI-3ARC Prize0.0%98.6%†
CritPtArtificial Analysis20.6%31.7%†
CritPt v2 hard subset (AA)Artificial Analysis–70.5%†
MysteryMechanismVals AI14.4%†53.1%†
Chess PuzzlesEpoch AI Benchmarking Hub43.0%†74.0%†
Mystery Game PuzzlesEpoch AI Benchmarking Hub21.0%†84.0%†
LMCAEpoch AI Benchmarking Hub48.5%64.1%
DTBenchEpoch AI Benchmarking Hub89.1%97.1%
BTF-3Epoch AI Benchmarking Hub–13.5%†
Coding
SciCodeSciCode50.5%56.5%†
WeirdMLWeirdML60.9%†93.3%†
FrontierCodeCognition39.8%†53.3%†
Show all 155 benchmarks
MetricPublisherGPT-5.6 Luna (xhigh)GPT-6 Astra
Coding
GSO-BenchGSO-Bench–79.4%
Arena CodingArena (LMArena)15021543†
Arena WebDevArena (LMArena)15191774†
LiveBench 2.0 CodingLiveBench–93.7%
LiveBench 1.0 CodingLiveBench82.9%†80.4%†
IOIVals AI61.8%†100.0%†
SWE-bench (Vals)Vals AI93.0%†–
Code MigrationVals AI44.5%†67.7%†
Code Migration v1.1Vals AI–80.6%†
ProgramBenchVals AI0.0%†5.5%†
SWE-bench MultimodalVals AI–43.3%†
Vibe Code Bench 1-100Vals AI22.6%†27.6%†
Vibe Code Bench v1.1Vals AI77.1%†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%†
Surface Evolver BenchEpoch AI Benchmarking Hub61.9%†–
DeepSWE v1.1Epoch AI Benchmarking Hub56.9%73.2%†
CursorBenchCursor33.0%–
ALE-BenchEpoch AI Benchmarking Hub1667.4†2951.3†
FrontierSWEEpoch AI Benchmarking Hub–65.5%†
Terminal-Bench 4.0 (AA)Artificial Analysis3.5%59.1%†
Terminal-Bench 2.1 (AA)Artificial Analysis77.9%88.4%†
Startrise FrontendStartrise65.0%†79.9%
FrontierCode 1.1 (AA)Artificial Analysis–47.0%†
FrontierCode 2.0 (AA)Artificial Analysis–42.4%†
Agents & tools
Terminal-BenchTerminal-Bench17.3%†58.2%†
Remote Labor IndexScale AI / CAIS–20.8%
APEX-AgentsMercor43.0%†64.7%
Arena AgentArena (LMArena)–14†
LiveBench Long AgenticLiveBench–77.5%
LiveBench BrowserLiveBench–91.3%
LiveBench Agentic CodingLiveBench48.4%†57.3%†
Terminal-Bench 2.1 (Vals)Vals AI79.0%†87.3%†
HiL-BenchScale AI SEAL–65.3%
GDPval-AA v2.1, normalised scoreArtificial Analysis–53.7%†
GDPval-AA v2.1Artificial Analysis–1584†
τ³-Banking (AA)Artificial Analysis28.7%41.4%†
ITBench SRE (AA)Artificial Analysis40.3%†48.6%†
Analyst Agent (AA)Artificial Analysis–51.3%†
APEX-Agents (AA)Artificial Analysis35.8%†–
CUA-benchVals AI–19.2%†
CyberBenchVals AI73.6%†41.1%†
Finance Agent v2Vals AI55.0%†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 AI1.3%†5.4%†
Legal Research BenchVals AI36.5%†39.4%†
Robo-benchVals AI–23.3%†
SkillsBenchVals AI60.5%†–
SREBenchVals AI1.5%†56.9%†
Tax Agent BenchVals AI24.8%†20.7%†
Terminal-Bench 4.0 (Vals)Vals AI11.6%†59.6%†
Terminal-Bench ScienceVals AI0.0%†62.9%†
Time Horizon Index: KSPVals AI–90.5%†
Arena SearchArena (LMArena)–1246†
BALROGBALROG45.6%†68.3%†
Vending-Bench 2Andon Labs4094.7†15514.7
AutomationBenchArtificial Analysis42.6%68.5%†
GDP.pdf (AA)Artificial Analysis23.8%31.0%†
Harvey LAB-AA v1.1Harvey–8.6%†
EnterpriseOps-GymArtificial Analysis40.8%†–
AA-Briefcase v1.1Artificial Analysis1339†1568†
Terminal-Bench-Science 0.1 (AA)Artificial Analysis3.3%†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 SearchGoogle79.3%†89.9%
Maths
FrontierMath Tiers 1–3Epoch AI Benchmarking Hub82.1%†93.7%†
FrontierMath Tier 4Epoch AI Benchmarking Hub61.0%†97.6%†
OTIS Mock AIMEEpoch AI Benchmarking Hub98.3%†100.0%†
ProofBenchVals AI60.0%†99.0%
LiveBench MathematicsLiveBench87.2%†96.8%†
Arena MathsArena (LMArena)14781487†
Knowledge
SimpleQA VerifiedEpoch AI Benchmarking Hub41.0%†75.6%†
LiveBench 2.0 Data AnalysisLiveBench–83.0%
LiveBench ResearchLiveBench–85.2%
LiveBench 1.0 Data AnalysisLiveBench78.0%†83.0%†
AA-Omniscience v1.1Artificial Analysis–42.4†
MMLU-ProVals AI86.0%†–
LegalBenchVals AI84.0%†–
CorpFinVals AI64.2%†–
TaxEvalVals AI76.2%†–
PRBench FinanceScale AI SEAL–47.5%
PRBench LegalScale AI SEAL–48.4%
BioMysteryBenchVals AI61.5%†79.3%†
Excel Modeling BenchmarkVals AI67.1%†71.7%†
Excel Modeling Benchmark v1.1Vals AI–72.7%†
MedCodeVals AI42.4%†48.5%†
MedScribeVals AI84.4%†87.9%†
Public Benefits Bench v1.1Vals AI61.2%†–
DrugDiscoveryBenchScale AI SEAL–68.7%†
EBR-benchEpoch AI Benchmarking Hub–76.2%†
GDP.pdfEpoch AI Benchmarking Hub22.7%†34.2%†
AA-Omniscience v1.1: accuracyArtificial Analysis–61.6%†
AA-Omniscience v1.1: non-hallucinationArtificial Analysis–50.2%†
HealthBench Professional (AA)OpenAI–66.0%†
Instruction following
LiveBench LanguageLiveBench72.6%†89.4%†
LiveBench Instruction FollowingLiveBench60.1%†75.6%†
Arena Instruction FollowingArena (LMArena)14441473†
Human preference
Arena TextArena (LMArena)14531477†
EQ-Bench 4EQ-Bench1146†–
Arena Creative WritingArena (LMArena)14101450†
Arena Multi-turnArena (LMArena)14591485†
Arena Longer QueriesArena (LMArena)14521489†
Multimodal
Arena VisionArena (LMArena)12571280†
LiveBench DesignLiveBench–82.0%
MMMU-ProArtificial Analysis78.5%86.9%†
MMMU-Pro (Vals)Vals AI85.0%†–
MortgageTaxVals AI67.3%†–
Blueprint-Bench 2Epoch AI Benchmarking Hub22.6%†49.7%
Code2Video-AAArtificial Analysis–1427†
AA-3DArtificial Analysis–1164†
Long context
AA-LCR v1.1Artificial Analysis81.7%80.7%†
Arena DocumentArena (LMArena)14621473†
MLCRArtificial Analysis15.0%41.7%†
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub156.5†168.8
LiveBench 2.0LiveBench–85.6%
LiveBench 1.0LiveBench73.6%†82.2%†
AA Intelligence Index v5.0Artificial Analysis–56.9†
AA Intelligence Index v4.3.2Artificial Analysis34.652.7†
AA Progress Index v1Artificial Analysis305.8349.6†
Vals IndexVals AI–65.0†
Vals Finance IndexVals AI–55.6%†
Vals RSI IndexVals AI–27.1%†
Vals Science IndexVals AI25.3%†65.1%†
Vals Multimodal IndexVals AI69.1%†–
Vals Router IndexVals AI–65.0%
AA Cyber IndexArtificial Analysis–33.4†

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

What changed

  • GPT-5.6 Luna (xhigh): 1 new result (AA Progress Index v1).
  • GPT-6 Astra: 3 score revisions.
  • GPT-6 Astra: 1 new result (Vals Router Index).
  • GPT-5.6 Luna (xhigh): 1 score revision.
  • GPT-6 Astra: 1 score revision.
  • GPT-6 Astra: 8 new results (LiveBench 2.0, LiveBench Long Agentic, LiveBench 2.0 Reasoning…).
  • GPT-5.6 Luna (xhigh): 3 new results (AutomationBench, GDP.pdf (AA), MLCR).
  • GPT-6 Astra: 1 new result (Startrise Frontend).
  • GPT-6 Astra: BenchLeader Index 68.9 → 68.0.
  • GPT-5.6 Luna (xhigh): 1 score revision.

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

GPT-5.6 Luna vs GPT-6 Astra: questions

Is GPT-6 Astra better than GPT-5.6 Luna?
GPT-6 Astra leads on quality: 68.2 vs 58.5. The BenchLeader Index combines every independent quality benchmark; GPT-6 Astra is ahead overall as of 29 Oct 2026, but check the category scores for your use.
GPT-5.6 Luna vs GPT-6 Astra: which is cheaper, and will that change?
GPT-5.6 Luna is cheaper today: $0.450 against $20.00 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
Is GPT-5.6 Luna or GPT-6 Astra better for coding?
GPT-6 Astra scores higher in coding (83 vs 57 on the category index, where 50 is average).
Which is faster, GPT-5.6 Luna or GPT-6 Astra?
GPT-5.6 Luna streams 3×+ faster (124 vs 40 output tokens per second, Artificial Analysis 7-day median).
Is GPT-5.6 Luna or GPT-6 Astra better for agents and tool use?
GPT-6 Astra scores higher in agentic tasks (70 vs 57 on the category index, where 50 is average).
More questions (4)
Which is better for long documents?
Both accept 1,050,000 tokens of context.
Which has lower latency?
GPT-6 Astra answers first: 3.0 s against 38.0 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.
Do independent benchmarks agree on GPT-5.6 Luna vs GPT-6 Astra?
Across the 2 benchmarks both have, GPT-5.6 Luna leads 0, GPT-6 Astra leads 2. Every overall index that rates both puts GPT-6 Astra ahead.
When was GPT-6 Astra released?
GPT-6 Astra was released on 2 Sept 2026.