# Qwen Image2.1 GGUF in ComfyUI doesn't work — retained record

> Retained reddit record from reddit:StableDiffusion.

Canonical: https://highsignal.app/data/records/0685b42046541791

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:StableDiffusion)  published 2026-09-21 15:01 ingested 2026-09-22 02:38

[original source ↗](https://www.reddit.com/r/StableDiffusion/comments/1wmf9jt/qwen_image21_gguf_in_comfyui_doesnt_work/)

## retained text



```
Pretty simple: I downloaded qwen2.1 Q4KM+vae+text encoders

Using comfyui Qwen2.1 workflow

Had to replace the load diffusion with a Unet Loader

The model is spitting out pure random noise and there is nothing I can do about it

Anyone able to use it?

EDIT: works with qwen3vl text encoder, not with Qwen3.5-9b one
```



document fetched 2026-09-22 00:31 · canonical https://reddit.com/r/StableDiffusion/comments/1wmf9jt/qwen_image21_gguf_in_comfyui_doesnt_work
