Can Qwen Image Edit 2509 maintain character consistency?

Yes. Qwen Image Edit 2509 supports character consistency across edits. To maintain the same character appearance in different scenes, keep the same seed and settings across runs.

Qwen Image Edit 2509 ComfyUI Guide: AI Image Editing (2026)

Qwen Image Edit 2509 is a powerful AI-powered image editing model for ComfyUI. It lets you edit photos using text prompts — changing backgrounds, swapping outfits, removing or adding objects, merging multiple characters into one scene, and altering facial expressions. This guide covers everything from model setup to real example runs.

What Qwen Image Edit Can Do

Qwen Image Edit 2509 enables advanced image editing with just a text prompt. Below are six editing types with real before/after examples — all generated from the three input images shown first.

Input Images Used for All Examples

Qwen Image Edit input 1 — anime girl with pink hair in casual outfit — Input 1

Qwen Image Edit input 2 — anime girl with dark hair in school uniform — Input 2

Qwen Image Edit input 3 — anime girl with light hair in dress holding basket — Input 3

Image Grouping / Multi-Image Merge

Combine multiple characters into a single coherent scene with natural lighting and proportions.

Qwen Image Edit — input image of single anime girl before multi-image merge — Input — click to enlarge

Qwen Image Edit output — three anime girls merged sitting together on a bench naturally — Output — click to enlarge

Prompt Used

Make the three girls sit together on a bench, maintaining natural proportions and lighting.

Qwen Image Edit — input image before side-by-side character merge — Input — click to enlarge

Qwen Image Edit output — three anime girls standing side-by-side with matched lighting and colors — Output — click to enlarge

Prompt Used

Combine the three girls into one photo. Place them standing side-by-side naturally and match the lighting and colors so they look like they were photographed together.

Clothing Replacement

Swap outfits while preserving pose, hairstyle, face, and art style.

Qwen Image Edit output — anime girl with outfit changed to bikini top and matching shorts — Output — click to enlarge

Prompt Used

Change her outfit to a cute bikini top with matching shorts. Keep her body pose, hairstyle, face, and overall style the same. Make the new outfit look natural and well-fitted with clean anime shading.

Object Removal

Remove objects from scenes and fill the area naturally without disturbing the rest of the image.

Qwen Image Edit output — basket removed from girl's hand with natural fill in watercolor style — Output — click to enlarge

Prompt Used

Remove the basket from her hand and fill the area naturally. Keep her pose, dress, and the soft watercolor style exactly the same.

Object Addition

Add new objects or characters to an existing image while matching lighting and art style.

Qwen Image Edit — anime girl before cat addition — Input — click to enlarge

Qwen Image Edit output — small cute cat added to girl's arms with matched anime style lighting — Output — click to enlarge

Prompt Used

Add a small cute cat that she is gently holding in her arms. Keep her pose mostly the same and make the cat fit naturally with the lighting and clean anime style.

Face Editing — Expression Change

Change facial expressions while keeping all other character features identical.

Qwen Image Edit — anime girl smiling before expression change — Input — click to enlarge

Qwen Image Edit output — anime girl with calm neutral expression replacing the previous smile — Output — click to enlarge

Prompt Used

Change her facial expression to a calm, neutral look. No smile, no happy emotion — just a relaxed, natural expression while keeping her face and style the same.

Background Change

Swap the background of any image while keeping the subject perfectly consistent.

Qwen Image Edit output — same anime girl now placed in a bright clean school hallway background — Output — click to enlarge

Prompt Used

Change the background to a bright, clean school hallway. Keep the girl exactly the same and match the lighting so she fits naturally into the new scene.

What Is Qwen Image Edit 2509?

Quick answer: Qwen Image Edit 2509 is an AI image editing model for ComfyUI. It accepts up to three input images and a text prompt, and outputs an edited image — handling anything from outfit changes to multi-character merges.

Qwen Image Edit 2509 is a powerful AI-powered image editing model designed for ComfyUI users. It allows you to edit photos, create consistent characters in different settings, apply new outfits, or transform backgrounds — all while maintaining exceptional output quality. The workflow supports creative possibilities, making it ideal for artists, content creators, and anyone interested in visual storytelling.

Download Workflow

Start by downloading the ready-made Qwen Image Edit 2509 workflow for ComfyUI. This file is pre-configured with all the correct node connections.

💡 Tip: Load the workflow into ComfyUI first before placing model files. It will show you exactly which node needs which file.

To use: drag and drop the JSON file into your ComfyUI workspace.

Download Required Models for Qwen Image Edit

All five required model files must be downloaded and placed in the correct folders for Qwen Image Edit to work. Download them from these official links:

Diffusion Model: qwen_image_edit_2509_fp8_e4m3fn.safetensors
GGUF: Qwen-Image-Edit-2509-XX_X.gguf
LoRA: Qwen-Image-Lightning-4steps-V1.0.safetensors
Text Encoder: qwen_2.5_vl_7b_fp8_scaled.safetensors
VAE: qwen_image_vae.safetensors

Folder Structure — Where to Place Model Files

Organize your model files in ComfyUI like this:

ComfyUI/
└── models/
    ├── diffusion_models/
    │   └── qwen_image_edit_2509_fp8_e4m3fn.safetensors
    ├── unet/
    │   └── Qwen-Image-Edit-2509-XX_X.gguf
    ├── loras/
    │   └── Qwen-Image-Lightning-4steps-V1.0.safetensors
    ├── vae/
    │   └── qwen_image_vae.safetensors
    └── text_encoders/
        └── qwen_2.5_vl_7b_fp8_scaled.safetensors

⚠️ Warning: Note the GGUF file goes into models/unet/ — not diffusion_models/. Placing it in the wrong folder is a common setup error.

Step-by-Step Setup: Qwen Image Edit 2509 in ComfyUI

Follow these nine steps to get Qwen Image Edit 2509 running inside ComfyUI on your local machine.

Update ComfyUI and Custom Nodes

Open ComfyUI Manager and click Update All to ensure all nodes are up to date before loading the new workflow. See how to install ComfyUI Manager →

Install the Workflow

Download the Qwen Edit 2509 workflow JSON file from the button above. Drag and drop it into your ComfyUI workspace to load it.

Qwen Image Edit 2509 ComfyUI workflow — node graph showing all connected model and image nodes — The Qwen Edit 2509 workflow loaded in ComfyUI

Install Missing Custom Nodes

If any nodes appear red after loading, open the Manager and click Install missing custom nodes to resolve them.

Download & Place All Required Models

Download all five model files listed above and place each file in its correct folder as shown in the folder structure section.

Assign Models in the Workflow

In each model node, select the corresponding file:

Load Diffusion Model: qwen_image_edit_2509_fp8_e4m3fn.safetensors
GGUF Loader: Qwen-Image-Edit-2509-XX_X.gguf
Load LoRA: Qwen-Image-Lightning-4steps-V1.0.safetensors
Load VAE: qwen_image_vae.safetensors
Load CLIP: qwen_2.5_vl_7b_fp8_scaled.safetensors

Load Your Image

Use the Load Image node(s) to import your input image(s). The workflow supports up to three input images. If you only want to edit one image, simply disable the other two input nodes.

ComfyUI Load Image nodes showing three input slots for Qwen Image Edit 2509 — two can be disabled for single-image editing — Load Image nodes — disable unused inputs for single-image editing

Edit Prompt and Settings

Add your creative editing prompt in the prompt node. Describe what you want to change — the more specific, the better. Adjust generation parameters like steps and CFG as needed.

(Optional) Enable Sage Attention

For faster generation, install Sage Attention. How to install Sage Attention → If not installed, simply disable the Sage Attention node — the workflow will still run normally.

Run the Workflow

Click Queue Prompt / Run to start editing your image with Qwen Image Edit 2509.

Example Run: Editing an Image with Qwen Image Edit 2509

Here's a complete example run showing input images, prompt, and output. All three input images shown above were used together.

Input Images

Example run input 1 — anime girl with pink hair — Input 1

Example run input 2 — anime girl with dark hair — Input 2

Example run input 3 — anime girl with basket — Input 3

Prompt Used

Combine the three girls into one photo. Place them standing side-by-side naturally and match the lighting and colors so they look like they were photographed together.

Output Result

Qwen Image Edit 2509 ComfyUI output — three anime girls merged standing side-by-side with consistent lighting and natural proportions — Output: three characters merged into one scene with matched lighting and colors

Tips & Troubleshooting

Missing nodes — red nodes in the workflow

Fix: Open ComfyUI Manager and click 'Install missing custom nodes'. Restart ComfyUI after installation. All nodes should turn green.

Model loading issues — file not found

Fix: Double-check every model file is correctly named and placed in the exact right folder. The GGUF file goes in models/unet/ specifically — not diffusion_models/.

Sage Attention errors

Fix: If Sage Attention errors appear, either install it using the guide linked in Step 8, or disable the Sage Attention node. The workflow will still run normally without it.

Unexpected output quality

Fix: If using alternate GGUF model sizes, adjust steps and CFG accordingly. Larger GGUF variants generally need fewer steps. Start with default settings and adjust from there.

💡 Tip: For consistent characters across multiple edits, keep the same seed value between runs. This significantly improves character consistency when making multiple different edits to the same person.

Conclusion

Qwen Image Edit 2509 offers a powerful and user-friendly way to perform advanced image editing in ComfyUI. Whether you want to alter backgrounds, maintain consistent characters across scenes, or explore creative prompt-driven transformations, Qwen Image Edit 2509 delivers high-quality results quickly. The support for up to three simultaneous input images makes it particularly powerful for character consistency and multi-image merging use cases.

Frequently Asked Questions

Qwen Image Edit 2509 is an AI-powered image editing model for ComfyUI. It accepts up to three input images and a text prompt, then outputs an edited image — supporting background changes, outfit swaps, object removal, multi-character merges, and expression changes.

You need five files: qwen_image_edit_2509_fp8_e4m3fn.safetensors (diffusion), a GGUF model, Qwen-Image-Lightning-4steps-V1.0.safetensors (LoRA), qwen_2.5_vl_7b_fp8_scaled.safetensors (text encoder), and qwen_image_vae.safetensors (VAE). All are free on Hugging Face.

Load up to three images using the Load Image nodes in the workflow. If you only need to edit one image, disable the other two input nodes. The workflow supports merging multiple images into one coherent scene.

Sage Attention is an optional component that speeds up generation. If it is not installed, you can disable the Sage Attention node and the workflow will still run normally, just slightly slower.

Yes. To maintain the same character appearance in different scenes or edits, keep the same seed and settings across runs. This significantly improves consistency when making multiple edits to the same character.