Qwen Image 2.1 LoRA image editing revises one to ten source images from a text instruction, with up to three of your own LoRA weights at strengths from 0 to 4. It returns one image at 1k, 1.5k or 2k.
Workbench
These examples were written for this endpoint. Replace the prompt and sample media URLs with your own accessible inputs, then check the current quote and model availability before submitting.
Outputpoolside-blue-hour With the Next Scene LoRA: Image 1 brings the man in his linen shirt, Image 2 the motel pool at blue hour, and the edit seats him on the pool's edge with his feet in the water.
- 1k
- 3:2
Outputcamisole-into-anime A photo turned into flat cel-shaded anime with an edit LoRA at scale 1.3. The instruction lists what must stay (the pose, the outfit, the mug, the window light) so only the rendering style changes.
- 2k
- 3:4
Outputnext-scene-sofa-night The Next Scene LoRA moves the same shot forward in time: hours later, the same couple asleep on the same sofa, the window gone dark and only the table lamp still on. The prompt repeats what each of them was wearing so the outfits don't change between scenes.
- 2k
- 3:2
Outputrelight-the-chaise A relight LoRA with the wardrobe and pose held fixed: candles from the right at mattress level, moonlight through the window behind. Both sources are new and the shadows are rebuilt around them.
- 2k
- 4:5
This example starts at $0.042 per image. The Playground updates the estimate when you change the settings.