MiniMax launches H3: a unified multimodal model for 2K video generation with native stereo audio

Short summary

MiniMax has released MiniMax H3, a general-purpose multimodal generation model that accepts text, images, video, and audio as unified input and outputs video with native stereo sound. Key specs include 2K resolution output and clip durations of 4–15 seconds. The model is positioned as a unified multimodal system rather than a text-to-video model with add-ons.

• •MiniMax H3 is a unified multimodal model accepting text, images, video, and audio as input
• •Outputs 2K video clips of 4–15 seconds with native stereo audio
• •Positioned as a general-purpose generation model, not just text-to-video

Generated with AI, which can make mistakes.