Model

Qwen2.5-Omni

Alibaba · open weights

qwenlm.github.io

Multimodal

Multimodal
BenchmarkScoreMetricHarnessSourceConfidenceLink
InteractionBench18.7scoreN/AIndependentmediumarxiv.org

Safety & Alignment

Safety & Alignment
BenchmarkScoreMetricHarnessSourceConfidenceLink
BASIS-Bench87.1demographic_robustness_—_content-based_tasks_(asr/sqa)_mean_accuracyN/AIndependentmediumarxiv.org
SEAR25.2t1_eer_↓N/AIndependentmediumarxiv.org

Reproduced by the community

Loading community runs…