On July 31, MiniMax announced the official release of MiniMax H3. According to the introduction, this is a general-purpose full-modal generation model that supports unified understanding of multi-modal contexts composed of text, images, video, and sound, and can output audio and video with native dual channels, and can support up to 15s 2K resolution. MiniMax said it plans to open up model weights within the next few days, provided that it complies with relevant laws and regulations.

On July 31, MiniMax announced the official release of MiniMax H3. According to the introduction, this is a general-purpose full-modal generation model that supports unified understanding of multi-modal contexts composed of text, images, video, and sound, and can output audio and video with native dual channels, and can support up to 15s 2K resolution. MiniMax said it plans to open up model weights within the next few days, provided that it complies with relevant laws and regulations.