General-purpose video model MiniMax H3 officially open-sourced, supports multimodal context and 2K resolution - IT Home
TL;DR AI
2 min readKey summary
MiniMax has released and open-sourced MiniMax H3, its next-generation general-purpose video model.
H3 supports multimodal understanding and generation across text, images, video, and audio, with multiple input modes and strong instruction following.
It can generate up to 15-second videos at up to 2K resolution, 24 FPS, and 32 kHz stereo audio.
The full workflow is covered by H3-Context-IR, H3-Base, and H3-Regenerate-2K for context retrieval, generation, and regeneration.
The launch pushes multimodal video generation forward and gives developers a ready-to-use open model and interface.
