True Index: Domain Rankings

Private Top-Tier (S & A) standing across 3 core domains
Agentic & Enterprise18.4p
#39
2 Top-Tier private tests
Coding & SWE26.3p
#35
3 Top-Tier private tests
Math & Scientific Reasoning34.2p
#34
2 Top-Tier private tests

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

3 matched domains · positive boost means higher legacy scores
Capability Domain Clean Evals (Tiers S & A) Legacy Benchmarks (Tiers C & D) Public Lift (Boost)
Agentic & Enterprise18.4p (2)21.2p (2)+2.9p
Coding & SWE26.3p (3)9.2p (1)-17.1p
General Knowledge & QA15.8p (3)9.6p (1)-6.2p
Non-comparable pre-release tests omitted from audited scoring (model released 2026-08-14): Terminal-Bench-Science 0.1 (2026-08-27).
Clean Evals Avg (Tiers S & A)
23.7p
10 private, leak-resistant tests
Legacy Benchmarks Avg (Tiers C & D)
13.3p
4 public or aging tests
Matched Public Lift (Boost)
-10.1p
Near zero or negative indicates clean generalization
Machine-Readable Export

Audited Benchmark Performance of Qwen3.8 27B

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#18/4665.0p
Code2Video-AAPrivate / FreshCoding & SWE#19/2322.6p
AA-Briefcase v1.1Private / FreshAgentic & Enterprise#39/5022.4p
AutomationBench-AAPrivate / FreshAgentic & Enterprise#43/5014.3p
FrontierCode-AA 2.0Private / FreshCoding & SWE#40/4614.2p
CyberBench v1.1Private / FreshCoding & SWE#17/2542.0p
LiveBenchPrivate / FreshGeneral Knowledge & QA#27/2820.6p
AA-Omniscience v1.1Private / FreshGeneral Knowledge & QA#40/4715.2p
ARC-AGI-2Private / FreshGeneral Knowledge & QA#22/2211.6p
CritPt v2Private / FreshMath & Scientific Reasoning#38/403.4p

Mid-Tier Evaluations (Tier B) (1)

Partially exposed or transitional evaluations with moderate leakage risk.

BenchmarkBenchmark Integrity StatusDomainRankPercentile
HLE-DiamondSemi-ExposedGeneral Knowledge & QA#42/4914.6p

Legacy & Vulnerable Benchmarks (Tiers C & D) (4)

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

BenchmarkBenchmark Integrity StatusDomainRankPercentile
GDP.pdfPublic / AgingAgentic & Enterprise#36/5028.6p
GDPval-AA v2.1Public / AgingAgentic & Enterprise#36/4313.9p
Terminal-Bench 4.0Public / AgingCoding & SWE#45/509.2p
MMMU-ProPublic / AgingGeneral Knowledge & QA#20/229.6p