Model

Qwen3-Omni

Alibaba · open weights

qwenlm.github.io

Multimodal

Multimodal
BenchmarkScoreMetricHarnessSourceConfidenceLink
BEANS-Next59.8accuracyN/AIndependentmediumarxiv.org
CogEmo-Bench52.2m-f1N/AIndependentmediumarxiv.org
OmniCapBench37.5native_constraintsN/AIndependentmediumarxiv.org
ReTurn95.0accuracyN/AIndependentmediumarxiv.org

Safety & Alignment

Safety & Alignment
BenchmarkScoreMetricHarnessSourceConfidenceLink
BASIS-Bench93.4demographic_robustness_—_content-based_tasks_(asr/sqa)_mean_accuracyN/AIndependentmediumarxiv.org
CARES39.0accuracyN/AIndependentmediumarxiv.org

Reproduced by the community

Loading community runs…