← sources

Give a try to Minimax H3-image ! (Crazy diversity)

source: reddit (reddit:StableDiffusion)published 2026-11-19 12:51ingested 2026-11-20 02:37

retained text

I think this model is heavily under-estimated, yes it have some flaws, but it has its strengths too.

My favorite trio for image generations are now Krea 2 (still the king to me with loras) / Moss / H3-image

The strength of this model is it's diversity/speed/realism and not as "sloppy" as some models.  
But it can also create some body horror (not flux3 level) and some stuff not making sense.

Because of the diversity it has, you really get this **Dopamine** effect when pressing the "generate" button. The same prompt can look vastly different without the cost of speed.

The image is an out of the box model comparaison, no loras, no nodes so you get an idea of the diversity.  
(All images took approx the same time to generate, between 8-13sec)

Prompt in comment.

EDIT :   
If you are planning to try the model, please use the SKILLs below. so you get better results.

Text to image : [https://github.com/MiniMax-AI/MiniMax-H3-Image/blob/main/skills/h3-image-prompt-writing/references/t2i-en.txt](https://github.com/MiniMax-AI/MiniMax-H3-Image/blob/main/skills/h3-image-prompt-writing/references/t2i-en.txt)

Image Edit : [https://github.com/MiniMax-AI/MiniMax-H3-Image/blob/main/skills/h3-image-prompt-writing/references/edit-en.txt](https://github.com/MiniMax-AI/MiniMax-H3-Image/blob/main/skills/h3-image-prompt-writing/references/edit-en.txt)

document fetched 2026-11-20 00:30 · canonical https://reddit.com/r/StableDiffusion/comments/1y08kz8/give_a_try_to_minimax_h3image_crazy_diversity