Alibaba releases Qwen3.5-Omni as a proprietary model, breaking from its open-source strategy

Alibaba released Qwen3.5-Omni, a multimodal AI model capable of processing text, audio, images, and video, but unlike its predecessors, the company is offering it as a proprietary model rather than open-source. The shift comes amid significant internal changes, including the departure of top AI researcher Junyang Lin, who helped establish Alibaba as a global leader in open-source AI, and a major reorganization that consolidated all of Alibaba's AI teams under one division led by CEO Eddie Wu. The new model can process up to 10 hours of continuous audio, supports 113 languages and dialects, and is positioned for enterprise use cases like video captioning, live-stream management, and customer service voice assistants.