TL;DR: This is a lightweight ComfyUI workflow designed to run Qwen Image on low to mid-end hardware.
It uses GGUF quantized models and optional Lightning LoRAs to drastically reduce VRAM usage while keeping image quality high.
✔ Tested on 12GB VRAM (Q8)
✔ Tested on 4GB VRAM (Q4_K_S)
✔ Supports both image generation and image editing
✔ Optimized for stability, speed, and low memory usage
Model links
You can find all the models on [QuantStack/Qwen-Image-Edit-2509-GGUF](https://huggingface.co/QuantStack/Qwen-Image-Edit-2509-GGUF) and [QuantStack/Qwen-Image-GGUF](https://huggingface.co/QuantStack/Qwen-Image-GGUF)
Diffusion model
- Choose your quant from the links above
LoRA
- [Qwen-Image-Lightning-4steps-V2.0.safetensors](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-4steps-V2.0.safetensors)
- [Qwen-Image-Lightning-8steps-V2.0.safetensors](https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-8steps-V2.0.safetensors)
Text encoder
- [qwen_2.5_vl_7b_fp8_scaled.safetensors](https://huggingface.co/Comfy-Org/Qwen-Image_ComfyUI/resolve/main/split_files/text_encoders/qwen_2.5_vl_7b_fp8_scaled.safetensors)
VAE
- [qwen_image_vae.safetensors](https://huggingface.co/Comfy-Org/Qwen-Image_ComfyUI/resolve/main/split_files/vae/qwen_image_vae.safetensors)
Model Storage Location
```
📂 ComfyUI/
├── 📂 models/
│ ├── 📂 diffusion_models/
│ │ └── Qwen-Image-Edit-2509-Q5_0.gguf
│ ├── 📂 loras/
│ │ └── Qwen-Image-Lightning-4steps-V2.0.safetensors
│ │ └── Qwen-Image-Lightning-8steps-V2.0.safetensors
│ ├── 📂 vae/
│ │ └── qwen_image_vae.safetensors
│ └── 📂 text_encoders/
│ └── qwen_2.5_vl_7b_fp8_scaled.safetensors
```
Custom Nodes
- [Kijay Nodes](https://github.com/kijai/ComfyUI-KJNodes)
-Editing mode: activate all nodes except the image uploads youre not using and connect the latent to the k sampler.
-Image mode: Connect the emptylatent node to the k sampler
If you have any questions feel free to ask check out my youtube channel if you want to see more :) : https://www.youtube.com/@the_frizzy1