MiniMax has a new model called H3, and it's positioning itself as a one-stop shop for video generation and editing. H3 can take in text, images, video, and audio at the same time, then follow a single natural-language instruction to create or edit video. For example, you could tell H3 to copy the camera movement from one clip, use a person from a different photo, and reference the sound from a third audio track. Previously, that kind of work would mean juggling separate tools for motion transfer, character reference, voice reference, and video editing — now it can happen in one task.
H3 can generate up to 15 seconds of video at 2K resolution with native stereo sound. Pricing for the official API is $0.13 per second for 2K output, which puts a max-length 15-second clip at about $1.95.