MiniMax releases open-weights H3 video model with native stereo audio and 2K output, day-0 support in ComfyUI · cho.sh
Sign in to create alerts.
MiniMax releases open-weights H3 video model with native stereo audio and 2K output, day-0 support in ComfyUI
AI video model releases
MiniMax releases open-weights H3 video model with native stereo audio and 2K output, day-0 support in ComfyUI
MiniMax H3 is MiniMax's third-generation video model and the first the company has released with open weights, following the closed Hailuo 01 and Hailuo 02.
H3 generates video up to 2K resolution and up to 15 seconds per clip, with audio produced in the same generation pass as native stereo rather than added afterward.
The model supports text-to-video, image-to-video, first-and-last-frame control, and reference-to-video (using reference images, video, or audio to carry a subject, motion, or voice through a clip).
H3 has multimodal context understanding, letting a single prompt combine images, audio, and video inputs and resolve their relationship into one output instead of requiring separate tasks.
ComfyUI added native support for H3 on release day, and the article states the model can run locally on a consumer RTX 3060 GPU.
cho.sh