MiniMax H3 generates 2K video with native stereo audio for a third of mainstream per-second pricing
- MiniMax is launching MiniMax H3, a general-purpose multimodal generation model that reads text, images, video and audio in one context and outputs video with native stereo sound, up to 15 seconds at 2K resolution.
- At 2K, H3's per-second price is less than a third of mainstream models, and at 768p it costs less than half of what mainstream models charge for 720p, with 2K as the default resolution.
- MiniMax plans to release H3 model weights as open in the coming days, subject to applicable laws and regulations, saying hardware compatibility was a design goal from the earliest stages.
- H3 is built on Contextual Omni Representation, H3-VAE, an H3-Omni Transformer and In-Context Regeneration, and MiniMax positions it for advertising, e-commerce, product design, UI/UX and gaming work.
- A demo prompt mixing a Hitchcock camera move from one video, a character from an image and vocals from a separate audio clip is handled as one request, and the company calls instruction following, text and brand rendering and V2V motion transfer strong points.