Qwen3-Omni is a natively end-to-end omni-modal LLM from Alibaba Cloud's Qwen team that understands text, audio, images, and video while generating speech in real time. Supporting 119 text languages, 19 speech input languages, and 10 speech output languages, it delivers state-of-the-art performance across multimodal benchmarks. The model processes mixed multimodal inputs simultaneously and produces streaming responses, making it suitable for conversational AI, captioning, translation, and real-time applications.