MiniMax launches open-weight H3 video model, undercutting rivals on 2K generation cost
AI video model releases
MiniMax launches open-weight H3 video model, undercutting rivals on 2K generation cost
MiniMax released its H3 video-generation model, which processes text, images, video and audio and generates clips up to 15 seconds long in 2K resolution with native stereo sound.
H3 can edit existing footage and transfer motion from one video to another using instructions and reference material in different formats.
MiniMax plans to release H3's model weights within days, making it open-weight in a video-generation field where most prominent models stay proprietary.
MiniMax says generating 2K video with H3 costs less than one-third of mainstream rival products, and the model is designed to run on several Chinese-made chips.
The release follows ByteDance's Seedance 2.0 and Kuaishou's Kling 3.0, intensifying competition among Chinese video-model developers; MiniMax is also developing a separate 2.7-trillion-parameter language model.
cho.sh