Native Active Perception as Reasoning for Omni-Modal Understanding

Benchmark Model Rank Results
audio-visual-question-answering-on-omnivideobenchOmniAgent-RL-7B#5Accuracy: 37.1
audio-visual-question-answering-on-worldsenseOmniAgent-RL-7B#12Accuracy: 47.2
audio-visual-question-answering-on-dailyomniOmniAgent-RL-7B#15Accuracy: 64.8
video-question-answering-on-lvbenchOmniAgent-RL-7B#35Accuracy: 50.5