Qwen-Image-Edit-2509 is an updated version of Alibaba’s image editing tool. The “2509” part means it came out in September 2025.
It’s built to let you tweak or mix images, follow your instructions, and keep things like faces or product designs consistent while making changes.
This version’s got a few big updates.
You can now edit 1 to 3 images together. That means you can throw in something like a person and a background and it’ll try to put them together in a natural way.
It holds on to key stuff better when editing one image. So if you’re changing a face or a product, it keeps details like look, texture or style more solid than before.
It can change the text inside pictures. You can switch words, change fonts or colors and it tries to keep things looking like they belong.
You get more control with extra guides. It works with outlines, depth maps, poses and other sketch-style inputs so you can steer edits better.
It’s open-source. Anyone can use or build on it under the Apache 2.0 license.
The model is based on Qwen-Image, which mixes AI image generation and transformer tech. That setup helps it edit images while keeping both meaning and looks pretty tight.
This version builds on that base with more training and extra tools.
People say it works best with up to 3 images. More than that and it can start to drop off. Tricky edits or really weird setups might still throw it off sometimes. It's great for restoring and recoloring old photos.
If you'd like to access this model, you can explore the following possibilities:
Version
A LoRA made to create image sequences that feel like they belong in a movie. It handles how the camera should move, how scenes should look, and makes sure shots connect well from one to the next.
Weights