Is MiniMax H3 Available in Diffusers?

Overview

Not as of early August 2026. Diffusers support requires released model weights, and MiniMax H3’s weights have not shipped yet — it was announced on July 31, 2026 as open-weight, but there is no public repository or model card, so there is nothing for the library to load. Source note: Verified 2026-08-06 against the MiniMax official H3 blog, MiniMax Video Generation API docs, and Hugging Face MiniMax-H3 model page.

Because of that, you cannot integrate H3 through Diffusers today; the only working access is the hosted API. Community discussion of the model exists, but that is not the same as an official released checkpoint or a Diffusers pipeline.

When the weights do land, library support may still lag the release and then arrive in a later version, so check the Diffusers documentation and the model card directly rather than assuming support from the model’s name. Pin whatever versions you rely on once support exists, so a routine update does not break your pipeline.

For now, if you are building tooling around H3, target the API. Re-check the Diffusers docs and MiniMax’s Hugging Face page after the weight release, since a new model’s ecosystem support arrives on its own timeline.