- Precise Text Editing: Qwen-Image-Edit supports bilingual (Chinese and English) text editing, allowing direct addition, deletion, and modification of text in images while preserving the original text size, font, and style.
- Dual Semantic/Appearance Editing: Qwen-Image-Edit supports not only low-level visual appearance editing (such as style transfer, addition, deletion, modification, etc.) but also high-level visual semantic editing (such as IP creation, object rotation, etc.).
- Strong Cross-Benchmark Performance: Evaluations on multiple public benchmarks show that Qwen-Image-Edit achieves SOTA in editing tasks, making it a powerful foundational model for image generation.
ComfyOrg Qwen-Image-Edit Live Stream
Qwen-Image-Edit ComfyUI Native Workflow Example
Make sure your ComfyUI is updated.Workflows in this guide can be found in the Workflow Templates.
If you can’t find them in the template, your ComfyUI may be outdated.(Desktop version’s update will delay sometime)If nodes are missing when loading a workflow, possible reasons:
- You are not using the latest ComfyUI version(Nightly version)
- You are using Stable or Desktop version (Latest changes may not be included)
- Some nodes failed to import at startup
1. Workflow File
After updating ComfyUI, you can find the workflow file from the templates, or drag the workflow below into ComfyUI to load it.
Download JSON Workflow
Download the image below as input
2. Model Download
All models can be found at Comfy-Org/Qwen-Image_ComfyUI or Comfy-Org/Qwen-Image-Edit_ComfyUI Diffusion model LoRA Text encoder VAE Model Storage Location3. Follow the Steps to Complete the Workflow

- Model Loading
- Ensure the
Load Diffusion Model
node loadsqwen_image_edit_fp8_e4m3fn.safetensors
- Ensure the
Load CLIP
node loadsqwen_2.5_vl_7b_fp8_scaled.safetensors
- Ensure the
Load VAE
node loadsqwen_image_vae.safetensors
- Ensure the
- Image Loading
- Ensure the
Load Image
node uploads the image to be edited
- Ensure the
- Prompt Setting
- Set the prompt in the
CLIP Text Encoder
node
- Set the prompt in the
- The Scale Image to Total Pixels node will scale your input image to a total of one million pixels,
- Mainly to avoid quality loss in output images caused by oversized input images such as 2048x2048
- If you are familiar with your input image size, you can bypass this node using
Ctrl+B
- If you want to use the 4-step Lighting LoRA to speed up image generation, you can select the
LoraLoaderModelOnly
node and pressCtrl+B
to enable it - For the
steps
andcfg
settings of the Ksampler node, we’ve added a note below the node where you can test the optimal parameter settings - Click the
Queue
button, or use the shortcutCtrl(cmd) + Enter
to run the workflow