Updated

MiniMax-H3-Image

MiniMax released MiniMax-H3-Image, an open-weight image generation and editing model built on the H3 family. It supports text-to-image, edits using reference images, and generation at resolutions up to 16 megapixels.

By No Code MBA · Facts checked

Provider
MiniMax
Announcement date
Status at last check
Generally available

What changed

MiniMax introduced a dedicated H3 image model that unifies text-to-image generation and image editing in one checkpoint. The company says the model supports up to 16 reference images and uses six sampling steps; it is a new image-generation release within the H3 family.

Access and availability

Download MiniMax-H3-Image weights from the official Hugging Face repository, or use the MiniMax API Platform and Hailuo AI. Local deployment instructions are provided in the model repository.

Official access documentation

What you can build

Test it on representative image-editing and high-resolution workflows, paying particular attention to reference-image fidelity and the hardware needed to run the weights.

Limitations

The main model repository lists approximately 68.65 GB for its BF16 diffusion-transformer weights and a separate MiniMax-H3 text encoder, so local use requires substantial storage and suitable accelerator hardware. MiniMax's speed and quality claims are provider-reported.

Sources and release history