Qwen-Omni-Turbo is a brand-new multimodal understanding and generation large model that supports text, image, speech, and video input comprehension, as well as mixed input understanding. It features simultaneous streaming generation for both text and speech, significantly enhanced multimodal content comprehension speed, and offers four natural conversational voice styles.This version is a snapshot version from March 26, 2025.