Qwen-Image 2.1: 7B Unified Generation and Editing with RGBA Output

What is Qwen-Image 2.1?

Qwen-Image 2.1 is the second generation of Alibaba's open-weight Qwen-Image line, and it replaces the 20B MMDiT backbone of the original release with a 7B visual generation component built from 32 single-stream DiT layers. The model card describes four changes:

• Compact and efficient. Mixed-granularity attention plus prefix KV cache reuse keeps image quality high at a much lower compute cost than the 20B generation.
• Native transparency with unified creation and editing. One model writes regular or transparent RGBA images from text, edits transparent layers, and extracts a subject out of a photograph.
• Versatile editing. Up to 10 reference images per request, local edits specified with circles, painted annotations or a separate mask, and identity preservation for people and products.
• Realistic textures and refined aesthetics. Improved typography, portrait lighting and fine detail.

Native output sizes start at 2048x2048 for 1:1 and go to 2400x1792 (4:3), 2528x1696 (3:2) and 2752x1536 (16:9), with the matching portrait ratios.

Transparent images with a real alpha channel

The headline change is transparency that survives into the file. Instead of generating on a flat backdrop and cutting the subject out afterwards with a matting model, Qwen-Image 2.1 writes the alpha channel itself, so stickers, product renders and UI assets come out ready to composite. The model card recommends an explicit prompt format to trigger it:

This is an RGBA image with transparency. <your subject description>.
The image has alpha channel and the background is transparent.

ComfyUI support

Qwen-Image 2.1 landed in ComfyUI on day one with the integration PR (Comfy-Org/ComfyUI #16400), and the official templates require ComfyUI 0.37.0 or newer. Repackaged single-file weights live in Comfy-Org/Qwen-Image-2.1, including BF16 and INT8 convrot transformers plus the Qwen3-VL-8B text encoder and optional prompt enhancer.