True Index: Domain Rankings

Private Top-Tier (S & A) standing across 3 core domains
Agentic & Enterprise44.1p
#29
3 Top-Tier private tests
Coding & SWE50.0p
#26
4 Top-Tier private tests
Math & Scientific Reasoning55.6p
#21
1 Top-Tier private test

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 & Enterprise44.1p (3)57.7p (2)+13.6p
Coding & SWE50.0p (4)63.8p (3)+13.8p
General Knowledge & QA48.8p (2)48.4p (1)-0.4p
Math & Scientific Reasoning55.6p (1)45.1p (2)-10.5p
Clean Evals Avg (Tiers S & A)
49.6p
10 private, leak-resistant tests
Legacy Benchmarks Avg (Tiers C & D)
53.8p
8 public or aging tests
Matched Public Lift (Boost)
+5.1p
Near zero or negative indicates clean generalization
Machine-Readable Export

Audited Benchmark Performance of Inkling 2 Small

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

Top-Tier Validated Tests (Tiers S & A) (10)

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

BenchmarkBenchmark Integrity StatusDomainRankPercentile
MLCR-AAPrivate / FreshMath & Scientific Reasoning#22/4655.6p
AA-Briefcase v1.1Private / FreshAgentic & Enterprise#23/5055.1p
FrontierCode-AA 2.0Private / FreshCoding & SWE#25/4648.1p
Code2Video-AAPrivate / FreshCoding & SWE#15/2337.3p
AutomationBench-AAPrivate / FreshAgentic & Enterprise#45/5010.2p
AA-AnalystAgentPrivate / FreshAgentic & Enterprise#9/2467.0p
CyberBench v1.1Private / FreshCoding & SWE#11/2558.3p
AA-3DPrivate / FreshCoding & SWE#13/2256.3p
AA-Omniscience v1.1Private / FreshGeneral Knowledge & QA#25/4749.1p
ARC-AGI-2Private / FreshGeneral Knowledge & QA#14/2248.4p

Mid-Tier Evaluations (Tier B) (2)

Partially exposed or transitional evaluations with moderate leakage risk.

BenchmarkBenchmark Integrity StatusDomainRankPercentile
HLE-DiamondSemi-ExposedGeneral Knowledge & QA#23/4954.2p
VIALS-AASemi-ExposedMath & Scientific Reasoning#15/2336.4p

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

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

BenchmarkBenchmark Integrity StatusDomainRankPercentile
SWE Atlas - Codebase QnAPublic / AgingCoding & SWE#4/1588.1p
SWE-bench Pro V2Public / AgingCoding & SWE#2/882.9p
GDP.pdfPublic / AgingAgentic & Enterprise#17/5066.3p
BioMysteryBenchPublic / AgingMath & Scientific Reasoning#13/2356.8p
GDPval-AA v2.1Public / AgingAgentic & Enterprise#21/4349.1p
Terminal-Bench-Science 0.1Public / AgingMath & Scientific Reasoning#29/4133.5p
Terminal-Bench 4.0Public / AgingCoding & SWE#40/5020.4p
MMMU-ProPublic / AgingGeneral Knowledge & QA#11/2248.4p