OpenModels
Qwen Image Edit, online
An editing model first: attach a picture, write what should be different, and it changes that and leaves the rest alone. Swap a garment, change the weather, remove something from the background. With nothing attached it falls back to plain text-to-image, so one selection covers both jobs.
Which plan: Included from Starter.
What it is good at
Edits by sentence
"Make the jacket leather, keep everything else" is a whole instruction — no masking step.
Leaves the rest alone
The parts you did not mention come back unchanged, which is the hard part of prompt editing.
Two jobs, one selection
Attach an image and it edits; attach nothing and it generates. No second model to pick.
How to prompt this model
Describe the finished picture without an image, or a precise change with one.
- With no reference attached, describe the finished picture: subject, composition, setting and light.
- For an edit, image 1 is the source. Identify what changes and preserve the identity, layout, wording and background details that must remain.
- Additional images provide reference context; explain each image's role using its attachment order.
Example prompt
Change the mug in image 1 to the blue glaze in image 2. Keep the mug's shape, handle, printed lettering, table and lighting unchanged.
Read the prompting tutorial. Select this model before using Enhance prompt to apply its guidance to your draft.
Sources: Qwen image editing guide.
The specifics
| Setting | Details |
|---|---|
| Made by | Alibaba |
| Runs | Prompt-driven editing, and text to image |
| Reference images | Up to 3 |
| Needs a picture? | No — text-only prompts work too |
| Opens at | Starter |
A prompt to start from
Change the season in this photograph to late autumn: warm fallen leaves on the ground, bare branches, low golden light. Keep the people, their clothing and the framing exactly as they are.
Open Qwen Image Edit in the studio — Choose the model and review your prompt and quote before generating.
Questions people ask
Do I need to mask the area I want changed?
No. Describe the change in words and the model works out what to touch; the rest of the frame is meant to come back unchanged.
How many pictures can I attach?
Up to three. The first one is the image being edited and sets the output shape; the others are references it can borrow from.
Can it generate an image from nothing?
Yes. With no picture attached it runs as a plain text-to-image model, so you do not need to switch models to start a new picture.
Compare models
- Nano Banana Pro
- GPT Image 2
- Seedream 4.5
- Seedream 5.0 Pro
- Seedance 2.0 Mini
- Kling 3.0 Pro
- Google Veo 3.1
- Grok 2 Imagine
Guides that use Qwen Image Edit
- How to edit a photo by describing the change — Free plan
- How to keep the same face across different scenes — Free plan
- How to prompt different image and video models — Free plan
Articles about Qwen Image Edit
- Qwen Image Edit vs Nano Banana Pro, for editing a photo — Model comparisons
- What AI images cost, per picture, on every model we run — What things cost
- Our own GPUs vs premium AI models, and when each one wins — How this works