Chinese AI firm MiniMax launched its H3 text-to-video model on July 31. The company's stock jumped over 13% following the announcement.
The multimodal model processes text, images, and audio to generate 15-second videos. It delivers 2K resolution with native stereo sound.
MiniMax priced the H3 model at one-third the cost of comparable flagship models. The model ranks first globally for video editing on the Artificial Analysis benchmark. The company plans to open-source the model's weights to foster broader adoption and developer customization.
The H3 model targets commercial applications in advertising, e-commerce, and gaming. This launch places MiniMax in direct competition with major AI players including ByteDance.