On February 10, 2026, Alibaba officially launched Qwen-Image-2.0, a cutting-edge image generation and editing model. Serving as the foundational image generation component for the Qwen large model suite, Qwen-Image-2.0 seamlessly blends image generation with editing functionalities. In the AI Arena text-to-image assessment, it achieved an impressive score of 1029 points, securing the third position globally. It outperformed models such as Seedream4.5 and Flux2-Max, coming in just behind Google's Nano Banana Pro and GPT Image1.5. The model can handle ultra-long text inputs of up to 1K tokens and produce high-resolution 2K outputs, accurately interpreting complex instructions to effortlessly generate professional-grade PPTs, infographics, and multi-panel comics. It particularly shines in Chinese text rendering, nearly flawlessly depicting illustrated ancient texts with hundreds of characters, like the Orchid Pavilion Preface. Qwen-Image-2.0 is an enhanced version built upon Qwen-Image and Qwen-Image-Edit, integrating image generation and editing within a single, streamlined architecture, resulting in substantial performance improvements for both tasks. Currently, the API is accessible for invite-only testing on Alibaba Cloud's BaiLian platform, and users can try it out at no cost through Qwen Chat.