What is Hailuo AI?
Hailuo AI is MiniMax's video platform. Its newest model, which MiniMax's documentation calls MiniMax H3 and we list as Hailuo 3, came out on July 31, 2026 according to the model release notes. The notes describe it as an open, general-purpose multimodal video model that reads text, images, video and audio.
It follows Hailuo 2.3 and 2.3 Fast, which MiniMax announced in October 2025. MiniMax has also released most of the H3 model weights under a community licence, for further development including fine-tuning.
What Hailuo 3 can do
MiniMax's video generation guide lists text-to-video, image-to-video from a first frame, a last frame or both, and generation from references, with output at 768P or 2K and lengths of 4 to 15 seconds. The launch post says it generates sound together with the picture, in stereo.
On MakeMidas, Hailuo 3 makes clips with sound at 768p or 2K, from text, a first frame, or a first and last frame.
Hailuo 3, Max, Max Fast and 2.3 Fast
MiniMax's pricing page lists H3 next to H3-Max and H3-Highspeed, which also come in 480P. Here, Hailuo 3 Max and Hailuo 3 Max Fast are the lower-cost options: both make silent clips from text, a first frame, or a first and last frame, at the resolutions in the table below.
Hailuo 2.3 Fast is the faster, lower-cost version of Hailuo 2.3, both announced in October 2025. Here it only animates a first frame you upload, making silent 6 or 10 second clips at 768p.













