Alibaba’s Qwen team has released Qwen-Image-2.1, an open-source image model that combines text-to-image generation and image editing in a single workflow. The model’s visual generation component has 7 billion parameters.
Qwen-Image-2.1 natively supports transparent image generation and editing, local edits, and composition using up to 10 reference images. Qwen said the model is designed to balance image quality, inference efficiency and cost.
The model supports native 2K output across multiple aspect ratios and is available through the Qwen repository, Hugging Face and ModelScope. [Qwen]