Alibaba releases Qwen3-Omni-Flash full-modal large model

Alibaba has released its next-generation multimodal large-scale model, Qwen3-Omni-Flash. The model supports seamless input and real-time streaming of text, images, audio, and video, significantly improving the stability of multi-turn dialogues and the naturalness of speech generation. Its system prompting and control capabilities have achieved a leap forward, allowing users to finely define specific persona styles such as "sweet girl" or "mature woman."