AI BriefingMiniMax Releases H3, Available in ComfyUI
Key point
MiniMax has released H3, an open-weight model that generates video and stereo audio simultaneously.
Details
MiniMax's H3 is an open-weight general-purpose multimodal video generation model supporting text, image, video, and audio.
The following features are available in ComfyUI:
- Text-to-video
- Image-to-video
- Generation based on first and last frames
- Reference-based video generation
It generates synchronized stereo audio alongside the video without separate synthesis, supporting dialogue, sound effects, ambient noise, and music.
The open-weight checkpoint generates videos up to 768p·15 seconds, while the MiniMax hosted version supports up to 2K resolution. It utilizes high-compression video representation to integrate multiple generation tasks into a single architecture and provides local configuration via ComfyUI.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Involution and Overseas Expansion of Chinese AI: Structural Limitations and US Response Strategies