benchgap
OpenAI · model

GPT-5.6 Luna benchmark scores

As of 2026-11-22, GPT-5.6 Luna (OpenAI) has measured scores on 62 benchmarks and estimated scores on 132 more.

BenchmarkScoreSource
AA-GPQA Diamond91.1%measured
AA-HLE39.5%measured
AA-SciCode53.6%measured
AIME2697.9%estimated ± 1.4 pp, medium confidence
Artificial Analysis Intelligence Index51.2%measured
CritPt20.6%measured
GDPval-AA48.0%measured
GPQA Diamond92.3%measured
HLE54.2%estimated ± 2.2 pp, high confidence
AA-LCR83.7%measured
LiveCodeBench v690.6%estimated ± 2.7 pp, high confidence
AA-Omniscience Accuracy42.7%measured
SciCode53.3%estimated ± 3.1 pp, high confidence
Terminal-Bench 2.184.7%measured
BrowseComp83.3%measured
HLE w/ tools57.0%estimated ± 4.1 pp, low confidence
LongBench v269.0%estimated ± 5.5 pp, low confidence
VITA-Bench45.8%estimated ± 6.7 pp, low confidence
τ²-bench results89.6%estimated ± 6.5 pp, low confidence
AA Coding Index71.5%measured
AA-MMMU-Pro78.6%measured
APEX-Agents35.5%estimated ± 4.9 pp, low confidence
APEX-Agents-AA35.8%measured
BioMysteryBench (human-difficult)48.2%estimated ± 6.8 pp, low confidence
DeepSearchQA92.2%estimated ± 4.5 pp, high confidence
SWE-bench Verified82.9%estimated ± 3.6 pp, high confidence
AutomationBench36.5%estimated ± 10.4 pp, low confidence
BFCL v476.5%estimated ± 2.7 pp, medium confidence
CyberGym77.9%measured
JobBench62.5%estimated ± 4.9 pp, medium confidence
SkillsBench55.0%estimated ± 4.1 pp, medium confidence
SWE-bench Pro62.7%measured
WideResearch82.8%estimated ± 1.6 pp, medium confidence
AA Agentic Index42.7%measured
MMLU-Pro85.2%estimated ± 4.4 pp, high confidence
FailureSensorIQ66.5%estimated ± 1.7 pp, high confidence
GPQA92.3%measured
MMLU90.5%estimated ± 0.9 pp, medium confidence
Vals GPQA Diamond91.7%measured
Vals LiveCodeBench81.7%estimated ± 3.0 pp, high confidence
Vals MMLU-Pro86.0%measured
MMMLU90.9%estimated ± 0.7 pp, low confidence
MMMU85.2%estimated ± 0.8 pp, high confidence
Chess Puzzles43.0%measured
AA MMLU-Pro87.7%estimated ± 0.7 pp, high confidence
Gert Labs54.3%measured
AA-AnalystAgent41.7%estimated ± 5.5 pp, low confidence
AA Harvey LAB v1.092.7%estimated ± 1.2 pp, medium confidence
AA Tau3 Banking39.6%estimated ± 2.0 pp, medium confidence
AA Terminal-Bench 2.180.9%measured
ApprenticeBench7.0%measured
ARC-AGI-188.0%measured
ARC-AGI-259.5%measured
BabaIsBench63.0%measured
CursorBench 3.152.9%estimated ± 3.4 pp, medium confidence
CursorBench 3.261.1%measured
EBR-bench31.4%estimated ± 9.3 pp, medium confidence
Terminal-Bench 3.014.3%measured
FrontierCode 1.1 Main43.9%estimated ± 1.6 pp, medium confidence
FrontierMath v2 (Tier 4)58.5%measured
FrontierMath v2 (Tiers 1-3)78.6%measured
FrontierSWE v217.3%estimated ± 5.8 pp, low confidence
Furniture Assembly42.5%measured
MaxxBench46.1%measured
MazeBench0.0%measured
MirrorCode21.1%measured
Mystery Game Puzzles21.0%measured
NYT Connections69.4%measured
OfficeQA Pro50.8%estimated ± 6.8 pp, medium confidence
OSWorld-Verified75.1%estimated ± 2.3 pp, medium confidence
ParseBench-166.9%measured
Terminal-Bench 4.117.6%measured
Terminal-Bench Hard57.3%estimated ± 2.8 pp, medium confidence
Vals SWE-bench93.0%measured
Terminal-Bench 2.1 (Vals AI)79.0%measured
VulcanBench v385.5%measured
CWE-Bench-AA54.0%estimated ± 4.5 pp, medium confidence
AA Cyber Index25.8%estimated ± 9.2 pp, low confidence
DeepsecBench-AA19.4%estimated ± 8.2 pp, medium confidence
HealthBench Professional (length-adjusted, Artificial Analysis)61.5%estimated ± 4.3 pp, medium confidence
HealthBench Professional (raw, Artificial Analysis)65.9%estimated ± 0.8 pp, medium confidence
AA ITBench40.3%measured
Bug Hunt Bench23.3%estimated ± 7.9 pp, low confidence
Chartography (tools)82.5%estimated ± 4.1 pp, low confidence
CursorBench 4.035.9%measured
CWE-bench v152.2%estimated ± 4.5 pp, low confidence
DeepSWE67.2%measured
DRACO86.9%estimated ± 0.7 pp, low confidence
FrontierCode 2.027.1%estimated ± 4.7 pp, medium confidence
FrontierPhysics3.6%measured
HLE w/o tools41.8%estimated ± 4.3 pp, high confidence
OSWorld 2.045.6%measured
OSWorld 2.168.5%estimated ± 3.6 pp, high confidence
OSWorld 2.1 Offline53.8%estimated ± 1.9 pp, medium confidence
PostTrainBench v1.135.6%estimated ± 1.7 pp, low confidence
ProgramBench77.6%estimated ± 6.0 pp, low confidence
SpireClimb5.8%estimated ± 7.2 pp, medium confidence
StealthMark37.8%measured
SWE-bench Multimodal55.3%estimated ± 5.6 pp, low confidence
Terminal-Bench 4.024.4%estimated ± 4.5 pp, high confidence
Terminal-Bench-Science 0.10.0%estimated ± 10.7 pp, low confidence
Toolathlon-Verified78.3%estimated ± 1.0 pp, medium confidence
VulcanBench Frontier v484.2%measured
AA AutomationBench56.4%estimated ± 4.8 pp, medium confidence
GDP.pdf21.1%estimated ± 4.2 pp, high confidence
AA Terminal-Bench 4.017.2%estimated ± 6.2 pp, medium confidence
ARC-AGI-30.2%measured
ARC-AGI-3 Provider Adapter0.9%estimated ± 4.1 pp, high confidence
cursorBench4137.3%estimated ± 2.0 pp, medium confidence
CWE-bench49.7%estimated ± 6.2 pp, medium confidence
MKA35.7%estimated ± 3.1 pp, high confidence
SkillsBench official leaderboard (BenchFlow)54.1%estimated ± 6.1 pp, medium confidence
BinaryAudit35.8%estimated ± 12.1 pp, low confidence
CompileBench86.6%estimated ± 10.0 pp, low confidence
OTelBench23.6%estimated ± 4.1 pp, medium confidence
Vibe Code Bench58.0%estimated ± 10.0 pp, low confidence
FACTS Multimodal37.4%measured
Hermes Bench71.0%estimated ± 1.5 pp, medium confidence
Hermes Index37.2%estimated ± 3.1 pp, high confidence
ParseBench-270.2%measured
CharXiv86.3%estimated ± 5.6 pp, low confidence
CharXiv w/o tools82.5%estimated ± 1.2 pp, high confidence
ExploitBench33.2%measured
ExploitGym12.4%measured
AI2D_TEST94.0%estimated ± 2.0 pp, medium confidence
C-Eval93.2%estimated ± 1.2 pp, medium confidence
Claw-Eval68.6%estimated ± 6.4 pp, medium confidence
CountBench97.3%estimated ± 2.9 pp, medium confidence
DynaMath85.5%estimated ± 3.0 pp, medium confidence
ERQA68.4%estimated ± 3.1 pp, high confidence
HMMT Feb 2026100.0%estimated ± 1.5 pp, low confidence
HMMT Nov 202595.8%estimated ± 1.5 pp, low confidence
MathVision87.8%estimated ± 3.0 pp, high confidence
MCP Atlas74.2%estimated ± 5.8 pp, low confidence
MCP-Tasks78.3%estimated ± 7.0 pp, low confidence
MLVU (M-Avg)86.1%estimated ± 2.2 pp, high confidence
MMAnswerBench93.8%estimated ± 2.4 pp, low confidence
MMLU-Redux96.3%estimated ± 0.8 pp, low confidence
MMMU-Pro78.4%measured
NL2Repo56.8%estimated ± 4.7 pp, high confidence
PostTrainBench v140.5%estimated ± 3.6 pp, medium confidence
QwenClawBench93.2%estimated ± 6.0 pp, low confidence
RealWorldQA83.7%estimated ± 2.8 pp, high confidence
ScreenSpot Pro75.6%estimated ± 9.6 pp, low confidence
SimpleVQA66.8%estimated ± 7.5 pp, medium confidence
SuperGPQA78.5%estimated ± 8.5 pp, low confidence
Terminal-Bench 2.078.5%estimated ± 5.8 pp, low confidence
Toolathlon53.4%measured
V*100.0%estimated ± 7.5 pp, medium confidence
Video-MME (w/o subtitle)84.7%estimated ± 1.3 pp, high confidence
Video-MME (with subtitle)87.0%estimated ± 3.0 pp, high confidence
VideoMMMU85.6%estimated ± 2.4 pp, high confidence
We-Math82.4%estimated ± 7.4 pp, low confidence
AA LiveCodeBench99.7%estimated ± 2.6 pp, medium confidence
HealthBench Hard32.0%measured
LiveCodeBench Pro85.8%estimated ± 6.0 pp, low confidence
MedXpertQA (MM)71.1%estimated ± 2.4 pp, high confidence
MedXpertQA (Text)61.3%estimated ± 4.5 pp, medium confidence
MMLU-Pro (Arcee)87.6%estimated ± 3.4 pp, low confidence
React Native Evals82.1%estimated ± 3.2 pp, high confidence
ResearchClawBench20.5%estimated ± 1.9 pp, medium confidence
PostTrainBench31.0%estimated ± 3.3 pp, medium confidence
DAYJOB1.1%measured
DAYJOB Healthcare1.2%measured
Finance Agent v256.3%estimated ± 2.4 pp, medium confidence
BioMysteryBench (human-solvable)89.2%estimated ± 0.6 pp, medium confidence
FrontierCode 1.1 Extended55.1%measured
HealthBench (raw)51.3%estimated ± 0.7 pp, medium confidence
HealthBench (length-adjusted)55.4%estimated ± 0.4 pp, medium confidence
HealthBench Professional55.7%measured
HealthBench Professional (raw)62.8%estimated ± 1.7 pp, medium confidence
HLE-Verified31.0%estimated ± 2.0 pp, medium confidence
LABBench280.9%estimated ± 1.8 pp, low confidence
ScarfBench65.4%estimated ± 3.8 pp, medium confidence
BlenderWorld (text-only)70.1%estimated ± 2.8 pp, medium confidence
Cookie-Bench77.2%estimated ± 1.3 pp, medium confidence
STEMBench2.5%estimated ± 0.2 pp, medium confidence
SimpleQA49.3%estimated ± 4.3 pp, low confidence
Agents' Last Exam26.3%estimated ± 3.3 pp, medium confidence
BabyVision w/ Python88.8%estimated ± 3.8 pp, low confidence
OpenHarmony Bench v1.058.9%estimated ± 2.4 pp, high confidence
SEC-Bench Pro54.3%estimated ± 3.9 pp, high confidence
BabyVision61.7%estimated ± 9.1 pp, low confidence
MMVU77.4%estimated ± 2.7 pp, medium confidence
PerceptionBench59.9%estimated ± 4.9 pp, low confidence
MRCRv285.9%estimated ± 11.3 pp, low confidence
MMLongBench-Doc61.6%estimated ± 2.6 pp, low confidence
OCRBench V271.2%estimated ± 3.3 pp, high confidence
BioBasicsBench29.8%estimated ± 2.2 pp, low confidence
LifeVal74.7%estimated ± 1.6 pp, medium confidence
MMMU-Pro w/ Python79.5%measured
ExploitBench Internal Port0.7%estimated ± 1.4 pp, medium confidence
PinchBench90.4%estimated ± 3.6 pp, low confidence
SWE-Atlas Codebase QnA59.8%estimated ± 4.8 pp, high confidence