When Vision Speaks for Sound

Benchmark Model Rank Results
audio-visual-question-answering-on-worldsenseQwen3-Omni-30B + Thud#6Accuracy: 50.3
audio-visual-question-answering-on-dailyomniQwen3-Omni-30B + Thud#11Accuracy: 67.9
video-question-answering-on-lvbenchQwen3-Omni-30B + Thud#29Accuracy: 52.1