Model

Qwen3.8 Flash

Alibaba · open weights

qwenlm.github.io

Knowledge & Reasoning

Knowledge & Reasoning
BenchmarkScoreMetricHarnessSourceConfidenceLink
InsClaimBench76.6accuracyN/AIndependentmediumarxiv.org

Coding

Coding
BenchmarkScoreMetricHarnessSourceConfidenceLink
MetaKernelBench35.1paired_liftcute dslIndependentmediumarxiv.org

Agentic / Tool Use

Agentic / Tool Use
BenchmarkScoreMetricHarnessSourceConfidenceLink
BOTTLED0.3f1N/AIndependentmediumarxiv.org
BrickBench1048.0eloN/AIndependentmediumarxiv.org
EngiWorld16.1engiscoreN/AIndependentmediumarxiv.org
WorldAuditBench12.7success_ratevla–vlm auditorsIndependentmediumarxiv.org
WorldAuditBench8.5success_ratevlm auditorsIndependentmediumarxiv.org

Multimodal

Multimodal
BenchmarkScoreMetricHarnessSourceConfidenceLink
EMBER-Bench42.4accuracyN/AIndependentmediumarxiv.org
GDP.pdf16.6gdp.pdf_scoreN/AIndependentmediumsurgehq.ai
HSS30.9pass@1N/AIndependentmediumarxiv.org
MVGraph50.4accuracyN/AIndependentmediumarxiv.org
PlotGround-1k42.6accuracyN/AIndependentmediumarxiv.org

Reproduced by the community

Loading community runs…