priorsBuilder

OpenAI · released 2026-07-30

GPT-5.6 Luna

Measured on 96 benchmarks across 6 reasoning efforts.

General frontier

43.6

#24 of 32 models · best at max

Reasoning efforts in your mix

max43.6#74 of 128
xhigh40.1#86 of 128
high36.4#101 of 128
medium25.4#121 of 128
low20.4#124 of 128
none15.3#127 of 128

Your mix’s benchmarks

8 of 8 measured

reasoning

Humanity's Last Exam

587 results
  1. #86max of 58739.5%
  2. #105xhigh of 58737.0%
  3. #132high of 58733.4%
  4. #175medium of 58725.8%
  5. #209low of 58719.8%
  6. #359none of 5877.2%

Leader: Opus 5.5 · 61.4%

knowledge

AA-Omniscience

526 results
  1. #189max of 526-10.3
  2. #193xhigh of 526-10.8
  3. #203high of 526-12.0
  4. #209medium of 526-13.2
  5. #214low of 526-14.7
  6. #254none of 526-24.9

Leader: Opus 5.5 · 46.4

reasoning

ARC-AGI-3

57 results
  1. #49max of 570.2%
  2. #50low of 570.2%
  3. #51medium of 570.2%
  4. #52high of 570.1%
  5. #56xhigh of 570.0%

Leader: GPT-6 Astra · 99.9%

reasoning

BullshitBench v2

196 results
  1. #95max of 19641.0%
  2. #104low of 19637.0%

Leader: Opus 4.8 · 95.0%

coding

DeepSWE v1.1

70 results
  1. #21max of 7067.2%
  2. #38xhigh of 7056.9%
  3. #55high of 7044.2%
  4. #69medium of 7011.3%
  5. #70low of 701.5%

Leader: GPT-6 Astra · 74.1%

math

FrontierMath Tier 4 (v2)

70 results
  1. #19max of 7061.0%

Leader: GPT-6.1 Sol · 100.0%

reasoning

SimpleBench

97 results
  1. #57xhigh of 9746.8%

Leader: Opus 5.5 · 88.4%

agentic

Terminal-Bench 4.0

35 results
  1. #30max of 3517.3%

Leader: Opus 5.5 · 64.8%

All 96 benchmarks

show 88 more

agentic

AA-Briefcase v1.1

224 results
  1. #63max of 2241342
  2. #76xhigh of 2241268
  3. #92high of 2241170
  4. #128medium of 224935
  5. #155low of 224741
  6. #163none of 224706

Leader: Sonnet 5.5 · 1823

agentic

Agent Arena

50 results
  1. #35xhigh of 50-1.5%

Leader: Opus 5.5 · 15.5%

agentic

APEX-Agents-AA

33 results
  1. #7max of 3335.8%

Leader: Gemini 3.5 Flash · 47.1%

agentic

AutomationBench (Zapier)

115 results
  1. #43max of 11517.1%
  2. #61xhigh of 11512.9%
  3. #75high of 1159.1%
  4. #96medium of 1154.3%
  5. #106low of 1151.8%

Leader: Gemini 4 · 51.3%

agentic

AutomationBench-AA

219 results
  1. #75max of 21950.2%
  2. #88xhigh of 21942.6%
  3. #102high of 21935.6%
  4. #124medium of 21923.1%
  5. #146low of 21911.7%
  6. #160none of 2195.0%

Leader: Gemini 4 · 77.5%

agentic

BioMysteryBench

27 results
  1. #25max of 2761.5%

Leader: Sonnet 5.5 · 81.1%

agentic

EnterpriseOps-Gym-AA

50 results
  1. #26max of 5040.8%

Leader: Fable 5 · 51.1%

agentic

Harvey's Legal Agent Benchmark

79 results
  1. #55max of 791.3%

Leader: Muse Spark 1.2 · 25.4%

agentic

ITBench-AA

45 results
  1. #18max of 4540.3%

Leader: GPT-5.6 Sol · 56.2%

agentic

Legal Research Bench

78 results
  1. #39max of 7836.5%

Leader: Muse Spark 1.3 · 55.3%

agentic

SkillsBench

34 results
  1. #6max of 3460.4%

Leader: DeepSeek V4.1 Flash · 69.8%

agentic

Tax Agent Bench

69 results
  1. #37max of 6924.8%

Leader: Fable 5.1 · 49.2%

agentic

Terminal-Bench 2.1

18 results
  1. #13max of 1875.7%

Leader: GPT-6 Astra · 87.4%

agentic

Terminal-Bench 4.0 (Vals)

47 results
  1. #33max of 4711.6%

Leader: Opus 5.5 · 65.2%

agentic

Terminal-Bench Science (Vals)

40 results
  1. #38max of 400.0%

Leader: GPT-6 Astra · 62.9%

agentic

Vending-Bench 2

67 results
  1. #40max of 67$4,095

Leader: GPT-6 Astra · $15,515

agentic

WeirdML v3

17 results
  1. #17xhigh of 175.2%

Leader: GPT-6 Astra · 42.2%

agentic

τ³-Banking

206 results
  1. #62max of 20631.1%
  2. #72xhigh of 20628.7%
  3. #82high of 20625.2%
  4. #106medium of 20617.7%
  5. #131low of 20612.8%
  6. #149none of 2069.7%

Leader: Qwen 3.8 Max · 51.3%

coding

Arena Image-to-WebDev

58 results
  1. #37xhigh of 581517

Leader: Opus 5.5 · 1749

coding

Arena WebDev

122 results
  1. #48xhigh of 1221519

Leader: Opus 5.5 · 1813

coding

Code Migration

77 results
  1. #22max of 7744.5%

Leader: Sonnet 5.5 · 69.8%

coding

FrontierCode 1.1 Extended

130 results
  1. #63max of 13055.1%
  2. #69xhigh of 13054.5%
  3. #88high of 13050.5%
  4. #107medium of 13041.2%
  5. #123low of 13029.0%

Leader: Opus 5.5 · 65.3%

coding

FrontierCode 1.1 Main

130 results
  1. #72max of 13039.8%
  2. #74xhigh of 13038.9%
  3. #86high of 13035.9%
  4. #107medium of 13025.7%
  5. #122low of 13015.4%

Leader: Opus 5.5 · 54.6%

coding

IOI

44 results
  1. #16max of 4461.8%

Leader: Gemini 4 · 100.0%

coding

ProgramBench v1 Almost Resolved

62 results
  1. #31max of 629.0%

Leader: Opus 5.5 · 65.0%

coding

ProgramBench v1 Fully Resolved

62 results
  1. #56max of 620.0%

Leader: Opus 5.5 · 18.5%

coding

ProgramBench v1 Raw Pass Rate

62 results
  1. #27max of 6268.3%

Leader: Opus 5.5 · 87.0%

coding

SciCode

226 results
  1. #73max of 22653.6%
  2. #90high of 22651.6%
  3. #101xhigh of 22650.5%
  4. #131medium of 22646.8%
  5. #135low of 22646.1%
  6. #165none of 22640.4%

Leader: Opus 5.5 · 66.9%

coding

SWE-bench Verified (Vals)

85 results
  1. #9max of 8593.0%

Leader: Opus 5 · 97.0%

coding

Terminal-Bench 2.1 (Vals)

74 results
  1. #11max of 7479.0%

Leader: Opus 5.5 · 87.6%

coding

Terminal-Bench Science-AA

46 results
  1. #36max of 463.3%

Leader: GPT-6 Astra · 63.3%

coding

Terminal-Bench v2.1-AA

238 results
  1. #41max of 23880.9%
  2. #54xhigh of 23877.9%
  3. #73high of 23869.7%
  4. #113medium of 23853.2%
  5. #134low of 23843.4%
  6. #144none of 23839.0%

Leader: Fable 5.1 · 91.4%

coding

Terminal-Bench v4-AA

216 results
  1. #83max of 21611.6%
  2. #103xhigh of 2163.5%
  3. #108high of 2162.5%
  4. #134none of 2161.0%
  5. #138medium of 2160.5%
  6. #161low of 2160.0%

Leader: Sonnet 5.5 · 63.6%

coding

Vibe Code Bench 1-100

22 results
  1. #5max of 2222.6%

Leader: Opus 5.5 · 30.4%

coding

Vibe Code Bench v1.1

105 results
  1. #31max of 10577.1%

Leader: Sonnet 5.5 · 92.4%

coding

WeirdML v2

127 results
  1. #42max of 12760.9%

Leader: GPT-6 Astra Pro · 93.6%

composite

AA Intelligence Index v4.3.2

610 results
  1. #79max of 61037.3
  2. #87xhigh of 61034.6
  3. #108high of 61032.1
  4. #161medium of 61025.0
  5. #212low of 61021.0
  6. #277none of 61015.5

Leader: Opus 5.5 · 57.6

composite

Arena Text

220 results
  1. #64xhigh of 2201452

Leader: Gemini 4 · 1525

composite

Arena Text — Multi-Turn

220 results
  1. #67xhigh of 2201457

Leader: Gemini 4 · 1553

composite

Vals Index

47 results
  1. #25max of 4751.7%

Leader: Gemini 4 · 68.9%

composite

Vals Multimodal Index v1.2

33 results
  1. #7max of 3369.1%

Leader: Fable 5 · 74.2%

cost

AA Intelligence Index Cost per Task

177 results
  1. #4low of 177$0.01
  2. #5none of 177$0.01
  3. #10medium of 177$0.02
  4. #21high of 177$0.04
  5. #35xhigh of 177$0.09
  6. #53max of 177$0.18

Leader: GPT-6 Luna · $0.00

cost

Cost to Run AA Intelligence Index

177 results
  1. #5none of 177$22.97
  2. #7low of 177$27.02
  3. #15medium of 177$44.32
  4. #25high of 177$108
  5. #40xhigh of 177$179
  6. #52max of 177$320

Leader: GPT-6 Luna · $10.63

creative

threejseval

34 results
  1. #27high of 341291

Leader: Opus 5.5 · 2079

creative

VoxelBench (text)

55 results
  1. #13max of 551856

Leader: GPT-6 Astra · 2732

economics

CorpFin (Vals)

117 results
  1. #40max of 11764.2%

Leader: Opus 5 · 73.2%

economics

Excel Modeling Benchmark

74 results
  1. #17max of 7467.1%

Leader: Fable 5.1 · 76.7%

economics

Finance Agent v2

78 results
  1. #22max of 7855.0%

Leader: Gemini 4 · 65.4%

economics

GDP.pdf

210 results
  1. #39max of 21024.0%
  2. #41xhigh of 21023.8%
  3. #49high of 21022.2%
  4. #94medium of 21015.2%
  5. #98low of 21014.2%
  6. #150none of 2106.0%

Leader: Muse Wonder Full (Max) · 32.6%

economics

GDPval-AA v2.1

288 results
  1. #56max of 2881459
  2. #72xhigh of 2881399
  3. #90high of 2881330
  4. #125medium of 2881129
  5. #162low of 288994
  6. #173none of 288919

Leader: Opus 5.5 · 1866

economics

MortgageTax (Vals)

84 results
  1. #26max of 8467.3%

Leader: Opus 5 · 72.1%

economics

TaxEval (Vals)

125 results
  1. #7max of 12576.2%

Leader: Muse Spark 1.2 · 80.4%

knowledge

AA-Omniscience Accuracy

526 results
  1. #104max of 52642.7%
  2. #105xhigh of 52642.4%
  3. #107high of 52641.8%
  4. #118medium of 52640.7%
  5. #124low of 52639.6%
  6. #191none of 52628.6%

Leader: Fable 5.1 · 67.2%

knowledge

AA-Omniscience Non-hallucination

526 results
  1. #238none of 52625.0%
  2. #401low of 52610.1%
  3. #415medium of 5269.1%
  4. #456high of 5267.6%
  5. #458xhigh of 5267.5%
  6. #460max of 5267.4%

Leader: MiniCPM5-1B (Non-reasoning) · 99.1%

knowledge

GPQA Diamond

539 results
  1. #51max of 53991.1%
  2. #74xhigh of 53989.5%
  3. #80high of 53989.2%
  4. #118medium of 53985.9%
  5. #156low of 53983.5%
  6. #336none of 53964.5%

Leader: GPT-6 Astra · 96.3%

knowledge

GPQA Diamond (Vals)

122 results
  1. #18max of 12291.7%

Leader: Gemini 3.1 Pro · 95.5%

knowledge

MedCode

94 results
  1. #48max of 9442.4%

Leader: Opus 5 · 63.6%

knowledge

MedScribe

96 results
  1. #34max of 9684.4%

Leader: Opus 5.5 · 91.4%

knowledge

MMLU Pro (Vals)

122 results
  1. #49max of 12286.0%

Leader: Fable 5.1 · 92.4%

knowledge

MMMU Pro (Vals)

81 results
  1. #25max of 8185.0%

Leader: Fable 5.1 · 90.6%

knowledge

MMMU-Pro

265 results
  1. #65max of 26578.6%
  2. #66xhigh of 26578.6%
  3. #77high of 26577.6%
  4. #91medium of 26575.8%
  5. #113low of 26574.0%
  6. #199none of 26560.3%

Leader: Opus 5.5 · 87.7%

knowledge

Public Benefits Bench v1.1

50 results
  1. #31max of 5061.2%

Leader: Opus 5 · 76.9%

long-context

AA-LCR v1.1

523 results
  1. #19max of 52383.7%
  2. #51xhigh of 52381.7%
  3. #74high of 52380.3%
  4. #154medium of 52375.0%
  5. #213low of 52370.0%
  6. #330none of 52342.7%

Leader: Kimi K3 · 88.7%

long-context

Arena Document

43 results
  1. #23xhigh of 431457

Leader: Opus 5 · 1516

long-context

MLCR-AA

94 results
  1. #33max of 9419.4%
  2. #44xhigh of 9416.1%
  3. #58high of 9412.2%
  4. #70medium of 948.3%
  5. #75low of 946.1%

Leader: Sonnet 5.5 · 75.0%

math

FrontierMath Tiers 1–3 (v2)

115 results
  1. #14max of 11582.1%
  2. #57low of 11541.4%
  3. #58none of 11539.6%

Leader: GPT-6.1 Sol · 93.7%

math

ProofBench v1.1

52 results
  1. #24max of 5260.0%

Leader: Sonnet 5.5 · 100.0%

other

Arena Vision

116 results
  1. #47xhigh of 1161258

Leader: Fable 5 · 1308

other

T3 Code usage share

85 results
  1. #9max of 851.5%

Leader: Opus 5.5 · 29.1%

reasoning

ARC-AGI-1

215 results
  1. #64max of 21590.7%
  2. #80xhigh of 21587.7%
  3. #110high of 21578.0%
  4. #145medium of 21556.5%
  5. #166low of 21539.5%
  6. #212none of 2153.3%

Leader: Fable 5 · 98.5%

reasoning

ARC-AGI-2

216 results
  1. #80max of 21659.6%
  2. #98xhigh of 21647.6%
  3. #110high of 21633.3%
  4. #145medium of 2167.4%
  5. #153low of 2165.1%
  6. #192none of 2160.8%

Leader: GPT-6 Astra · 95.0%

reasoning

Blueprint-Bench 2

28 results
  1. #25max of 2822.6%

Leader: Gemini 4 · 54.4%

reasoning

CritPt

529 results
  1. #57max of 52920.6%
  2. #58xhigh of 52920.6%
  3. #83high of 52916.6%
  4. #154medium of 5294.9%
  5. #186low of 5292.6%
  6. #277none of 5290.3%

Leader: Muse Wonder Full (Max) · 32.9%

reasoning

Furniture Assembly

31 results
  1. #14max of 3142.5%

Leader: Opus 5.5 · 83.3%

reasoning

Humanity's Sixth Sense

25 results
  1. #24max of 2521.6%

Leader: GPT-6 Astra · 53.6%

reasoning

LegalBench

135 results
  1. #42max of 13584.0%

Leader: Fable 5 · 88.6%

reasoning

MysteryMechanism

27 results
  1. #26max of 2714.4%

Leader: GPT-6 Astra · 53.2%

reasoning

Roboflow Visual Reasoning

102 results
  1. #49high of 10265.6%
  2. #62low of 10260.5%

Leader: GPT-6 Astra · 91.2%

reasoning

SAGE

80 results
  1. #45max of 8044.2%

Leader: Opus 4.7 · 56.1%

reasoning

SimpleQA Verified

47 results
  1. #22max of 4739.1%

Leader: Gemini 3.1 Pro · 77.5%

reliability

Tool-call error rate

171 results
  1. #65max of 1710.9%

Leader: GPT-5.4 Pro · 0.0%

reliability

Uptime (7d)

181 results
  1. #129max of 18198.6%

Leader: GPT-5 Pro · 100.0%

safety

CyberBench v1.1

45 results
  1. #11max of 4573.6%

Leader: Wonder · 80.4%

safety

SRE Bench

32 results
  1. #24max of 321.5%

Leader: GPT-6 Astra · 56.9%

speed

Latency (Time to First Token)

351 results
  1. #26none of 3510.7s
  2. #119low of 3511.6s
  3. #176medium of 3512.4s
  4. #297high of 35129s
  5. #311xhigh of 35145s
  6. #334max of 351117s

Leader: Gemini 2.5 Flash-Lite · 0.3s

speed

Output Speed (tokens/s)

351 results
  1. #120max of 351122.2
  2. #125xhigh of 351118.7
  3. #136high of 351111.9
  4. #137none of 351111.8
  5. #142medium of 351110.2
  6. #148low of 351106.9

Leader: Celeris-1 · 1522.5

speed

Provider latency

177 results
  1. #116max of 1771.8s

Leader: Qwen 2.5 7B · 0.2s

speed

Provider throughput

176 results
  1. #33max of 17698.0

Leader: Muse Spark 1.1 · 204.3