# I'm liking Qwen 2.1—Midjourney style. — retained record

> Retained reddit record from reddit:StableDiffusion.

Canonical: https://highsignal.app/data/records/2e827d2bf6fb496b

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


source: [reddit](https://highsignal.app/data/reddit) (reddit:StableDiffusion)  published 2026-10-03 11:01 ingested 2026-10-04 02:37

[original source ↗](https://www.reddit.com/r/StableDiffusion/comments/1wwkhd9/im_liking_qwen_21midjourney_style/)

## retained text



```
I’m using it now, and as you can see from the images I generated—using the standard workflow without any tweaks other than setting the resolution to 2.0MP and using no LoRA—it has a distinct Midjourney-like look. That’s something I’ve been looking for in open-source models for a long time; the only other one that had that Midjourney style was Chroma V48-dc.



I still plan to run several tests; my current go-to for image generation is Krea2. I really liked Qwen 2512 back in the day, but it became obsolete, so this Qwen 2.1 seems like a significant step forward. I’ll also be testing it for editing tasks, where I currently use Klein 9b.



Now I’m going to run various tests with different prompts, but I can already say that the visual style it produced in these images is incredible.
```



document fetched 2026-10-04 00:30 · canonical https://reddit.com/r/StableDiffusion/comments/1wwkhd9/im_liking_qwen_21midjourney_style
