Introducing P-H3-Image. Built by Pruna AI, powered by MiniMax H3 open weights. • Text-to-Image & Image Editing • Up to 4K resolution • Multi-image references • Up to 3 custom LoRAs per request • Fast 8-step generation Create, edit, and customize with P-H3-Image.

Try P-H3-Image

That's awesome—exactly what I wanted! I've been using the Wan 2.x model to generate images, but it's getting pretty old, and the resolution and detail aren't quite there. That said, it does have really good spatial awareness, human anatomy, and texture. I hope H3 has better visuals.

@PrunaAI Massive 🚀🚀

@PrunaAI Cool bro

@PrunaAI 4K output, multiple image references, and support for custom LoRAs in just 8 steps make this a powerful option for creators who want both speed and creative control. 🔥

@PrunaAI Is this local ?

@PrunaAI Three LoRAs in one request is huge for keeping a character and a style together ✨

@PrunaAI Please check your inbox

@PrunaAI Up to 4k resolution is too good

@PrunaAI Pruna's P-H3-Image efficiently converts MiniMax H3 video into fast open image model.

@PrunaAI Awesome

@PrunaAI this 4k look too real how u do it so fast

@PrunaAI Such high quality

@PrunaAI So, that's an illegal thing to do right?

@PrunaAI Wow that's Amazing 😍👏