Benchmark Almanac

GPT-5.6 Luna

GPT-5.6 Luna ranks first on CorpBench Work.

OpenAI Released July 9th, 2026 Compare with others
261
Benchmarks with a score
1
Ranked first, out of five or more models
10
In the top three
OpenAI Jul 2026

GPT-5.6 Luna

Closed weights

156.7 #35 of 256
Epoch Capabilities Index
Self-Reported Index 155.1 #59 of 404
  • MMMU-Pro 79.5% #18 /90
  • DeepSWE 67.2% #10 /28
  • AutomationBench 17.1% #20 /42

Where it ranks

Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.

Knowledge and language

BenchmarkScoreRankSource
Wer wird Millionär? LLM Benchmark 416667.0 #4 of 77 Official leaderboard
StudyArena Text 1,286 #7 of 85 Official leaderboard
Creative Writing v3 78.5 #19 of 120 Official leaderboard
HLE Extended 57.5% #5 of 31 Official leaderboard
LMArena (Elo) 1,453 #60 of 335 Official leaderboard
AA-LCR Extended 72.2% #11 of 49 Official leaderboard
Longform Creative Writing (EQ-Bench) 75.5 #27 of 119 Official leaderboard
Last Translation Benchmark 15.4 #10 of 36 Official leaderboard
SEA-HELM 72.1 #28 of 71 Official leaderboard
GraphWalks 81.3% #11 of 27 Lab model card
SimpleQA Verified 39.1% #24 of 56 Official leaderboard
OpenAI MRCR v2 45.1% #21 of 47 Official leaderboard
BullshitBench V2 41% #62 of 129 Official leaderboard
Document Arena 1,460 #21 of 39 Official leaderboard
BullshitBench 25.4% #61 of 107 Official leaderboard
Extended NYT Connections 69.4% #59 of 103 Official leaderboard
FACTS Parametric V2 50.6% #16 of 27 Official leaderboard
HUMAINE 25 #33 of 50 Official leaderboard
DesignArena ASCII Art 1,142 #52 of 77 Official leaderboard
FACTS Parametric 25% #31 of 45 Official leaderboard
Global MMLU Lite 83.1% #30 of 43 Official leaderboard
FACTS Search V2 79.3% #15 of 21 Official leaderboard
EQ-Bench 4 1,146 #23 of 32 Official leaderboard
FACTS Benchmark Suite 56.9% #19 of 23 Official leaderboard
Continual Learning Exam 25.3% #10 of 11 Official leaderboard
FACTS Grounding v2 52.5% #23 of 25 Official leaderboard
FACTS Multimodal 37.4% #22 of 23 Official leaderboard

Math and science

BenchmarkScoreRankSource
Verifying the Verifiers (Project APE CRED benchmark) 97.5% #3 of 23 Official leaderboard
GPQA Diamond 91.6% #33 of 229 Epoch AI run
OTIS Mock AIME 2024-2025 98.3% #31 of 212 Epoch AI run
ProofBench 54% #11 of 59 Official leaderboard
FrontierMath Tiers 1-3 v2 82.1% #18 of 86 Official leaderboard
FrontierMath Tier 4 v2 61% #17 of 68 Official leaderboard
DTBench 89.1% #57 of 180 Official leaderboard
ARC-AGI 90.7% #32 of 95 Official leaderboard
Surface Evolver Bench 61.9% #9 of 25 Official leaderboard
ARC-AGI-2 59.6% #39 of 98 Official leaderboard
ProofBench v1.1 60% #24 of 52 Official leaderboard
SimpleBench 46.8% #64 of 106 Official leaderboard
FrontierPhysics 3.60% #14 of 22 Official leaderboard
scBench (LatchBio SingleCellBench) 50.3% #21 of 26 Official leaderboard
BioMysteryBench 61.5% #26 of 30 Official leaderboard
LatchBio Capabilities v1.0 33.5% #24 of 27 Official leaderboard
MysteryMechanism 14.4% #30 of 31 Official leaderboard
Terminal-Bench-Science 3.33% #17 of 17 Official leaderboard

Coding

BenchmarkScoreRankSource
IOI (Vals) 72.9% #4 of 55 Official leaderboard
AndroidBench 87.6% #5 of 32 Official leaderboard
SWE-bench Verified Extended 74.7% #8 of 44 Official leaderboard
ALE-Bench 1,829 #31 of 126 Official leaderboard
WeirdML 60.9% #30 of 112 Official leaderboard
SWE-Bench Pro 62.7% #22 of 81 Lab model card
Code Migration 44.5% #22 of 78 Official leaderboard
Vibe Code Bench 1-100 22.6% #8 of 26 Official leaderboard
Vibe Code Bench 77.1% #41 of 120 Official leaderboard
SWE Atlas Codebase QnA Extended 19.3% #13 of 38 Official leaderboard
DeepSWE 67.2% #10 of 28 Official leaderboard
ReactBench 33.9% #10 of 28 Official leaderboard
WebDev Arena 1,519 #38 of 105 arena.ai, via Epoch AI
DesignArena Website 1,248 #57 of 151 Official leaderboard
IOI v2 (Vals) 61.8% #17 of 45 Official leaderboard
DesignArena UI Component 1,241 #56 of 139 Official leaderboard
DesignArena SVG 1,187 #43 of 103 Official leaderboard
Code Arena 1,519 #54 of 128 Official leaderboard
Brokk Power Ranking 8.80 #9 of 19 Official leaderboard
DesignArena Code 1,221 #68 of 143 Official leaderboard
DesignArena Game Dev 1,191 #70 of 143 Official leaderboard
SWE-sweep 2.50% #5 of 10 Official leaderboard
DesignArena Data Visualization 1,204 #65 of 118 Official leaderboard
SWE Refactor Bench 10.5% #5 of 8 Official leaderboard
DesignArena Agentic Overall 1,141 #47 of 74 Official leaderboard
Image-to-WebDev Arena 1,517 #35 of 55 Official leaderboard
Data Agent Benchmark 69.6% #12 of 18 Official leaderboard
DesignArena 3D Design 1,102 #94 of 136 Official leaderboard
SWE-rebench 43.6% #9 of 13 Official leaderboard
APEX-SWE 32.1% #50 of 65 Official leaderboard
SWE-Clar 49% #7 of 9 Official leaderboard
DesignArena Image to HTML 1,159 #41 of 51 Official leaderboard
WeirdML v3 5.16% #24 of 27 Official leaderboard
CursorBench 35.9% #19 of 21 cursor.com, via Epoch AI
CursorBench 4.0 35.9% #19 of 21 Official leaderboard
MirrorCode 21.1% #36 of 39 Epoch AI run
Chart Arena 1,366 #14 of 15 Official leaderboard
Vision2Web 64.5% #10 of 10 Official leaderboard
Artificial Analysis Coding Agent Index v1.1 74.6 #3 of 3 Lab model card

Agents and tools

BenchmarkScoreRankSource
PinchBench 2.0 88.7% #7 of 57 Official leaderboard
LLM Chess 1,185 #17 of 123 Official leaderboard
LisanBench 4180.7 #27 of 133 Official leaderboard
BALROG 45.6% #8 of 33 Official leaderboard
Toolathlon 53.4% #8 of 29 Lab model card
BrowseComp 83.3% #25 of 81 Lab model card
Game Arena Poker 4.80 #10 of 31 Official leaderboard
AppWorld Test-Challenge 73.4% #4 of 12 Official leaderboard
MyPCBench 55.4% #3 of 9 Official leaderboard
Chess Puzzles 43% #33 of 93 Epoch AI run
RuneBench 5.29 #28 of 76 Official leaderboard
Terminal-Bench 2.1 Extended 35.4% #18 of 44 Official leaderboard
GAIA 60.8% #28 of 64 Official leaderboard
AutomationBench 17.1% #20 of 42 Official leaderboard
GameHorizon 67.8 #18 of 37 Official leaderboard
AppWorld Test-Normal 85.1% #6 of 12 Official leaderboard
Mystery Game Puzzles 21% #47 of 84 Epoch AI run
Terminal-Bench 2.1 75.7% #8 of 13 Official leaderboard
Game Arena Chess 700 #26 of 42 Official leaderboard
o11y-bench 65.6% #14 of 22 Official leaderboard
Game Arena Lines of Action 321 #26 of 39 Official leaderboard
Brood War Bench 50% #4 of 6 Official leaderboard
SnakeBench 21 #126 of 188 Official leaderboard
Terminal-Bench 4.1 17.6% #29 of 43 Official leaderboard
Terminal-Bench 4.0 17.3% #27 of 40 Official leaderboard
BrowseComp Extended 16.5% #29 of 42 Official leaderboard
Cresta-Bench 38.8% #10 of 14 Official leaderboard
Game Arena Word Art 259 #28 of 39 Official leaderboard
OSWorld-Science 22.6% #8 of 11 Official leaderboard
Game Arena Dark Hex 144 #30 of 41 Official leaderboard
Game Arena Bargaining 156 #29 of 39 Official leaderboard
BabaIsBench 63% #24 of 32 Official leaderboard
BU Bench V2.1 69.6 #12 of 16 Official leaderboard
Era by Eon 56.6% #7 of 9 Official leaderboard
CLASHarness 19.8% #11 of 14 Official leaderboard
DuelLab GameBench 2 54.0 #45 of 57 Official leaderboard
Game Arena Word Association 82 #33 of 40 Official leaderboard
TextQuests 22.5 #24 of 28 Official leaderboard
Game Arena Four in a Row 187 #37 of 43 Official leaderboard
Game Arena Go 446 #34 of 39 Official leaderboard
Kaggle Game Arena 168 #37 of 42 Official leaderboard
ApprenticeBench 7.00% #16 of 18 Official leaderboard
OSWorld 2.0 45.6% #17 of 19 Lab model card
Game Arena Checkers 323 #37 of 41 Official leaderboard
Terminal-Bench 3.0 14.3% #10 of 11 Official leaderboard
Game Arena Clobber 37 #36 of 39 Official leaderboard
Game Arena Nine Men's Morris 33 #37 of 39 Official leaderboard
Game Arena Ultimate Tic-Tac-Toe 22 #37 of 39 Official leaderboard
Game Arena Dots and Boxes 502 #38 of 40 Official leaderboard
ARC-AGI-3 0.18% #26 of 27 Official leaderboard
Game Arena Coin Game 0 #39 of 39 Official leaderboard
Game Arena Reversi 0 #39 of 39 Official leaderboard
DuetBench 951 #9 of 9 Official leaderboard

Professional work

BenchmarkScoreRankSource
TaxEval v2 76.2% #6 of 129 Official leaderboard
CorpBench Work 92.7 #1 of 17 Official leaderboard
Vals Index v2 59.9% #13 of 65 Official leaderboard
ProfBench 55.3% #16 of 72 Official leaderboard
MMEnterprise-Bench 65.8% #2 of 9 Official leaderboard
Agents' Last Exam 30.3% #7 of 31 Official leaderboard
Excel Modeling Benchmark 67.1% #18 of 76 Official leaderboard
PolicyBench 88.7% #12 of 50 Official leaderboard
MortgageTax 67.3% #25 of 87 Official leaderboard
Vals Finance Agent v2 55% #24 of 80 Official leaderboard
APEX-1 66.6% #22 of 71 Official leaderboard
LegalBench 84% #43 of 138 Official leaderboard
CorpFin v2 64.2% #40 of 119 Official leaderboard
FoodTruck Bench $32,932 #21 of 59 Official leaderboard
MedScribe 84.4% #36 of 97 Official leaderboard
TRLawBench 9.25 #13 of 34 Official leaderboard
BenchCAD 73.9% #5 of 11 Lab model card
GDPval-AA v2 1,592 #13 of 27 Lab model card
MedXpertQA MM Extended 56.2% #20 of 41 Official leaderboard
Vals Index v2.1 51.7% #25 of 49 Official leaderboard
MedCode 42.4% #49 of 96 Official leaderboard
Legal Research Bench 36.5% #48 of 93 Official leaderboard
Tax Agent Bench 24.8% #48 of 84 Official leaderboard
Management Consulting Tasks (OpenAI) 35.4% #4 of 7 Official leaderboard
HealthBench Consensus 95.1% #10 of 17 Lab model card
Vending-Bench 2 $4,095 #41 of 67 Official leaderboard
HealthBench Hard 32% #13 of 21 Lab model card
Public Benefits Bench 61.2% #31 of 50 Official leaderboard
DAYJOB: Healthcare 1.60% #25 of 37 Official leaderboard
Tuesday Work Index 43.3 #30 of 44 Official leaderboard
Harvey LAB (Legal Agent Benchmark) 1.25% #56 of 80 Official leaderboard
DAYJOB: Finance 4.30% #28 of 39 Official leaderboard
APEX-Accounting 54.3% #36 of 49 Official leaderboard
CHI-Bench 13.3% #17 of 23 Official leaderboard
Vending-Bench Arena $2,434 #24 of 32 Official leaderboard
GDPval Extended 0.40% #33 of 43 Official leaderboard
APEX-Agents 1.1 57.8% #50 of 65 Official leaderboard
APEX-Agents 43% #47 of 61 Collected by Epoch AI, source not named
Genspark Slides Benchmark 54.1% #8 of 10 Official leaderboard
StrataBench Land v1 Extract 85.9% #4 of 5 Official leaderboard
HealthBench 55.8% #25 of 31 Lab model card
CommerceAgentBench 44.9% #11 of 13 Official leaderboard
YC-Bench $196,650 #52 of 59 Official leaderboard
HealthBench Professional 55.7% #23 of 26 Lab model card
RCM-Bench-1 71.5% #19 of 20 Official leaderboard
Agents' Last Exam v1.1 5.74% #14 of 14 Official leaderboard
BigFinanceBench 36% #2 of 2 Lab model card

Vision, audio and robotics

BenchmarkScoreRankSource
MMMU-Pro 79.5% #18 of 90 Lab model card
ZeroBench 21% #20 of 76 Official leaderboard
ExtractBench 91.2% #8 of 27 Official leaderboard
Vision Arena 1,255 #41 of 136 Official leaderboard
4DCodeBench 0.67 #5 of 16 Official leaderboard
MMMU-Pro Extended 39.2% #12 of 38 Official leaderboard
Chartography 31.4% #14 of 43 Official leaderboard
VoxelBench 1,817 #21 of 56 Official leaderboard
ParseBench 66.9% #19 of 44 Official leaderboard
vCX Benchmark 86.9% #21 of 42 Official leaderboard
BVB 66.7 #23 of 45 Official leaderboard
Furniture Assembly Benchmark 42.5% #18 of 35 Official leaderboard
CharXiv Extended 78.7% #23 of 38 Official leaderboard
RoTA-Bench 4.00% #5 of 8 Official leaderboard
ArtifactArena 984 #15 of 21 Official leaderboard
Blueprint-Bench 2 22.6% #30 of 38 Official leaderboard
GDP.pdf 22.7% #14 of 17 Lab model card
InteractiveBench 20% #11 of 12 Official leaderboard
Humanity's Sixth Sense 21.6% #33 of 34 Official leaderboard

Safety and security

BenchmarkScoreRankSource
CyberBench (Vals) 83.9% #2 of 26 Official leaderboard
SAFE-Teen Urgent safety 23% #2 of 17 Official leaderboard
Tacit Knowledge and Troubleshooting 83% #4 of 30 Lab model card
SAFE-Teen Health advice 21% #3 of 17 Official leaderboard
CyberBench v1.1 (Vals) 73.6% #10 of 44 Official leaderboard
SAFE-Teen Relationship boundaries 35% #4 of 17 Official leaderboard
GeneBench Pro 23.6% #3 of 12 Official leaderboard
First-Person Fairness evaluation 0.61% #3 of 10 Lab model card
Human Pathogen Capabilities Test v2 (HPCT v2) 62.5% #23 of 71 Official leaderboard
DeepsecBench 26.6 #10 of 29 Official leaderboard
nocot-bench (NCRI) 76.6 #70 of 185 Official leaderboard
DistressBench 72.7% #17 of 42 Official leaderboard
Agent JSK 99.9% #3 of 7 Lab model card
BioTIER Refuse 72.1% #31 of 69 Official leaderboard
TroubleshootingBench 37.7% #9 of 20 Lab model card
PCT (Polarized Contrastive Pairs) 44.7% #23 of 46 Official leaderboard
Capture the Flag (Internal) 93.2% #2 of 4 Lab model card
BioTIER Permit 99.3% #37 of 69 Official leaderboard
dfbench 52.4% #10 of 18 Official leaderboard
BioSecBench-Surveillance 43.2% #17 of 27 Official leaderboard
CVE-Bench 82.3% #9 of 14 Lab model card
Multimodal Troubleshooting Virology 50.2% #18 of 27 Lab model card
SpatialBench-Long 25% #18 of 27 Official leaderboard
Cyber Range 61.5% #7 of 10 Lab model card
scBench-Long 13.1% #19 of 27 Official leaderboard
Prompt injection robustness 97.1% #5 of 7 Lab model card
ProtocolQA Open-Ended 29.9% #20 of 27 Lab model card
SRE-Bench 1.53% #26 of 35 Official leaderboard
TxBench-Preclinical-Pharmacology 43.7% #23 of 30 Official leaderboard
EpiBench 19.8% #21 of 27 Official leaderboard
OncologyBench 10.7% #25 of 30 Official leaderboard
BioSecBench-Function 27.6% #23 of 27 Official leaderboard
ExploitBench 33.2% #18 of 21 Lab model card
ExploitGym 12.4% #13 of 15 Lab model card
VariantBench 27.1% #24 of 27 Official leaderboard
MetagenomicsBench 24.3% #24 of 26 Official leaderboard
SecIT Bench 63.8% #24 of 26 Official leaderboard
TxBench-Antibody-Discovery 19.3% #29 of 30 Official leaderboard
TxBench-Oligo-Discovery 23% #29 of 30 Official leaderboard
SEC-bench Pro 48.9% #11 of 11 Lab model card
Destructive action avoidance 73% #8 of 8 Lab model card
Instruction hierarchy attacks 99.9% #7 of 7 Lab model card
LifeSciBench 51.2% #5 of 5 Lab model card
MedChemBench (OpenAI) 30.4% #5 of 5 Lab model card
Prompt injection attacks in search and function calling 89.7% #3 of 3 Lab model card

Other

BenchmarkScoreRankSource
Vals Multimodal Index 69.1% #7 of 33 Official leaderboard
LMCA 48.5% #32 of 147 Official leaderboard
Conceptual Reasoning Index 58.7 #43 of 151 Official leaderboard
Political Science LLM Benchmark 67% #16 of 42 Official leaderboard
ACCoRD 67.5 #62 of 152 Official leaderboard
CAIS Vision Capabilities Index 53.7 #24 of 53 Official leaderboard
Xbench 36.7 #8 of 17 Official leaderboard
CAIS Risk Index 50.3 #33 of 67 Official leaderboard
JevBench 93.6% #2 of 4 Official leaderboard
CAIS Text Capabilities Index 31.3 #36 of 68 Official leaderboard
SAGE (Vals) 44.2% #45 of 81 Official leaderboard
Artificial Analysis Intelligence Index v4.1 51.2 #7 of 8 Lab model card
Vals Science Index 25.3% #24 of 25 Official leaderboard

Its lab's numbers put it at 155.1, 1.6 points below its ECI

The Self-Reported Index fits one score to the 18 capability results OpenAI reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 151.2 to 158.9. How we compute it.

Its lab's numbers next to 11 leaderboards

Where the benchmark's own leaderboard, or Epoch AI's run, also has this model. Gaps mostly come from different harnesses, effort levels and task subsets. Every pair.

Benchmark Lab says Board says Gap
Agents' Last Exam 50.3% Agents' Last Exam 30.3% Leaderboard, xhigh, codex +20 pts
GeneBench Pro 10.8% GeneBench Pro 23.6% Leaderboard, pro −13 pts
Terminal-Bench 2.1 84.7% Terminal-Bench 2.1 75.7% Leaderboard, max, Codex +9.0 pts
OpenAI MRCR v2 41.3% OpenAI MRCR v2 8-needle 256K-512K 41.3 (512K-1M 41.3) 45.1% Leaderboard, 8-needle, 128K to 256K bin, max reasoning −3.8 pts
FrontierMath Tiers 1-3 v2 78.6% FrontierMath Tier 1-3 (v2) 82.1% Leaderboard, max; task version 2.0.0, Epoch −3.5 pts
FrontierMath Tier 4 v2 58.5% FrontierMath Tier 4 (v2) 61% Leaderboard, max; task version 2.0.0, Epoch −2.5 pts
AutomationBench 14.9% AutomationBench 17.1% Leaderboard, max −2.2 pts
GPQA Diamond 92.3% GPQA Diamond 91.6% Epoch AI run, max +0.7 pts
DeepSWE 67.2% DeepSWE v1.1 67.2% Leaderboard, max, mini-swe-agent Same
ARC-AGI-3 0.18% ARC-AGI-3 0.18% Leaderboard, max, standard Same
Management Consulting Tasks (OpenAI) 35.4% Management Consulting Tasks (Internal) 35.4% Leaderboard Same

OpenAI's models over time

OpenAI's models in color, against every other model on the Epoch Capabilities Index, by release date.

50 100 150 2023 2024 2025 2026 2027 GPT-6.1 Terra GPT-6 Luna GPT-5.4 Mini GPT-5.2 Pro GPT-5 GPT-5 nano GPT-4.1 nano 50 100 150 2023 2024 2025 2026 2027 GPT-6.1 Terra GPT-6 Luna

Source: Epoch Capabilities Index, Epoch AI (CC BY 4.0)

Its launch post

The benchmarks the lab chose to headline, with the scores it gave.

GPT-5.6 (OpenAI) Jul 9, 2026

Its system cards

Every result in the longer documents the lab published with it. 30 of them aren't in the launch post.

GPT-5.6 Preview System Card Jun 25, 2026

Health

Hallucination

Safety

Robustness

Bias

Bio and chem risk

Cyber risk

AI R&D risk

GPT-5.6 System Card Jul 9, 2026

Health

Hallucination

Safety

Robustness

Bias

Bio and chem risk

Cyber risk

AI R&D risk

GPT-5.6 – August Updates Aug 6, 2026

Health

Hallucination

Safety

Robustness

Bio and chem risk

Cyber risk