True Index: Domain Rankings

Private Top-Tier (S & A) standing across 3 core domains
Agentic & Enterprise34.3p
#35
5 Top-Tier private tests
Coding & SWE50.3p
#25
5 Top-Tier private tests
Math & Scientific Reasoning57.1p
#20
3 Top-Tier private tests

Head-to-Head: Clean Evals vs. Legacy Benchmarks

4 matched domains · positive boost means higher legacy scores
Capability Domain Clean Evals (Tiers S & A) Legacy Benchmarks (Tiers C & D) Public Lift (Boost)
Agentic & Enterprise34.3p (5)37.8p (2)+3.5p
Coding & SWE50.3p (5)23.5p (2)-26.8p
General Knowledge & QA22.0p (3)35.5p (1)+13.5p
Math & Scientific Reasoning57.1p (3)39.7p (3)-17.5p
Clean Evals Avg (Tiers S & A)
40.9p
16 private, leak-resistant tests
Legacy Benchmarks Avg (Tiers C & D)
34.1p
8 public or aging tests
Matched Public Lift (Boost)
-10.5p
Near zero or negative indicates clean generalization
Machine-Readable Export

Audited Benchmark Performance of GPT-6 Luna

Grouped by benchmark integrity so private capability signal stands apart from aging public tests.

Top-Tier Validated Tests (Tiers S & A) (16)

Private, leak-resistant tests that build real capability signal.

BenchmarkBenchmark Integrity StatusDomainRankPercentile
MLCR-AAPrivate / FreshMath & Scientific Reasoning#16/4669.3p
FrontierMathPrivate / FreshMath & Scientific Reasoning#15/2758.4p
Harvey LAB-AA v1.1Private / FreshAgentic & Enterprise#15/2152.3p
FrontierCode-AA 2.0Private / FreshCoding & SWE#26/4645.8p
Code2Video-AAPrivate / FreshCoding & SWE#13/2344.6p
MKA (Multimodal Knowledge Agent)Private / FreshAgentic & Enterprise#18/2338.2p
AutomationBench-AAPrivate / FreshAgentic & Enterprise#39/5022.4p
AA-Briefcase v1.1Private / FreshAgentic & Enterprise#41/5018.4p
CyberBench v1.1Private / FreshCoding & SWE#4/2580.7p
AA-3DPrivate / FreshCoding & SWE#14/2252.4p
CritPt v2Private / FreshMath & Scientific Reasoning#21/4043.7p
AA-AnalystAgentPrivate / FreshAgentic & Enterprise#17/2440.4p
LiveBenchPrivate / FreshGeneral Knowledge & QA#24/2830.7p
Code Migration v1.1Private / FreshCoding & SWE#18/2427.9p
ARC-AGI-2Private / FreshGeneral Knowledge & QA#19/2225.4p
AA-Omniscience v1.1Private / FreshGeneral Knowledge & QA#43/479.8p

Mid-Tier Evaluations (Tier B) (2)

Partially exposed or transitional evaluations with moderate leakage risk.

BenchmarkBenchmark Integrity StatusDomainRankPercentile
HLE-DiamondSemi-ExposedGeneral Knowledge & QA#31/4937.5p
VIALS-AASemi-ExposedMath & Scientific Reasoning#22/234.5p

Legacy & Vulnerable Benchmarks (Tiers C & D) (8)

Public or aging benchmarks prone to overfitting and training-data leaks.

BenchmarkBenchmark Integrity StatusDomainRankPercentile
HealthBenchPublic / AgingMath & Scientific Reasoning#10/1661.0p
GDP.pdfPublic / AgingAgentic & Enterprise#21/5057.1p
Terminal-Bench-Science 0.1Public / AgingMath & Scientific Reasoning#28/4137.2p
SWE Atlas - Codebase QnAPublic / AgingCoding & SWE#14/1528.6p
BioMysteryBenchPublic / AgingMath & Scientific Reasoning#22/2320.8p
GDPval-AA v2.1Public / AgingAgentic & Enterprise#34/4318.5p
Terminal-Bench 4.0Public / AgingCoding & SWE#41/5018.4p
MMMU-ProPublic / AgingGeneral Knowledge & QA#14/2235.5p