China's MiniMax unveils AI video model 70% cheaper than rivals
What's the story
Chinese artificial intelligence (AI) firm MiniMax has launched a new video-generation model, the H3. The Shanghai-based company's latest innovation can process text, images, videos, and audio. The H3 model is capable of generating videos of up to 15 seconds in 2K resolution with native stereo sound. It can also edit existing content and transfer movements between different videos based on instructions and reference material provided in various formats.
Open-source initiative
Model weights to be released soon
MiniMax plans to release the model weights of its H3 system in the coming days.
This would allow users to download and customize the underlying system, expanding the open-weight approach already adopted by Chinese AI developers into video generation.
The move comes as many leading models in this space still remain proprietary.
Market focus
MiniMax went public in Hong Kong earlier this year
MiniMax, which was founded in 2022, is one of China's "AI tigers," a group of well-funded start-ups competing with domestic tech giants and US leaders in the foundation models market.
The company went public in Hong Kong earlier this year.
The H3 model is targeted at commercial applications such as advertising, e-commerce, product design, and gaming.
MiniMax claims that generating 2K video with its new model would cost less than a third of mainstream rival products.