# Qwen Image 2.1 - 1girl examples — retained record

> Retained reddit record from reddit:StableDiffusion.

Canonical: https://highsignal.app/data/records/6db0fdf1f917067f

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


source: [reddit](https://highsignal.app/data/reddit) (reddit:StableDiffusion)  published 2026-09-19 19:50 ingested 2026-09-20 02:37

[original source ↗](https://www.reddit.com/r/StableDiffusion/comments/1wkvqlj/qwen_image_21_1girl_examples/)

## retained text



```
Hey guys! So, I got early access to Qwen Image 2.1, and I did what any sane person would do, generate 200+ "1girl" images (while testing keywords and prompt structures). Here are some amateur-style images that I generated. If you want me to run any of your prompts, just put them in the comments.  
  
My thoughts -> I think the images are quite good, especially given that it is a 7B parameter model, but the output does come out a bit "noisy." The editing and reference image capabilities look good as well.
```



document fetched 2026-09-20 00:30 · canonical https://reddit.com/r/StableDiffusion/comments/1wkvqlj/qwen_image_21_1girl_examples
