Minimax H3

MiniMax H3 is an AI video generator that accepts text, images, video and audio together as inputs to produce finished 2K clips with matching stereo audio. You can provide character portraits, guide clips and soundtrack samples so the output keeps a consistent look, motion and sonic style.

It supports long-prompt text-to-video, dual-keyframe image animation and video motion transfer, and renders common aspect ratios including 9:16, 16:9 and 21:9. The system outputs native 2K footage with 32kHz stereo audio so there’s less need for separate audio syncing or post-production stitching.

• Mixed-modal inputs: text, images, video and audio
• Text-to-video, keyframe animation and motion transfer
• Native 2K output with synchronized 32kHz stereo audio
• Supports 9:16, 16:9 and 21:9, renders crisp text and logos