Hailuo, MiniMax H3, H3 Max and Turbo

MiniMax’s Hailuo family now includes the base H3 model and fal’s post-trained H3 Max and H3 Max Turbo variants. Max and Turbo have separate benchmark runs here. Their launch prices end September 30, 2026; compare both the promotional and regular rates.

MiniMax · facts checked

Which one to pick

Side by side, measured

ModelPriceSeconds to make one clip9 clips costLength
H3 Max Turbo$0.02 per second at 768p
via fal
6 to 6$0.905 to 15 seconds
H3 Max$0.04 per second at 768p
via fal
6 to 26$1.805 to 15 seconds
MiniMax H3$0.06 per second at 768p
via fal
52 to 182$2.705 to 15 seconds
Hailuo 2.3$0.28 per six-second clip, standard 768p
via fal
98 to 197$2.526 or 10 seconds, nothing in between; our five-second fixture rounds to six

Seconds and cost are from our own runs of the same 9 prompts through each model; open a model for every clip and the request behind it.

Every version

VersionReleasedWhat changed
video-01MiniMax's first text-to-video model under the Hailuo name.
Hailuo 021080p, better physics and instruction following.
Hailuo 2.3
hailuo-2.3
More natural motion and faces, stronger stylisation, a Fast variant at up to half the cost.
MiniMax H3
minimax-h3
One transformer for text, image, video and audio in; video with native stereo sound out, up to 2K and 15 seconds; base weights open-sourced 3 August 2026.
H3 Max
minimax/h3-max/text-to-video
Date not verifiedfal post-trains MiniMax H3 for prompt adherence and faster video generation. Max supports text and image inputs, native audio and clips of 5 to 15 seconds.
H3 Max Turbo
minimax/h3-max-turbo/text-to-video
Date not verifiedH3 Max Turbo is fal’s lower-priced post-trained MiniMax H3 variant. It shares the 5 to 15 second envelope and resolution choices of Max; this page tests what the cheaper tier returns.

Frequently asked questions

Is Hailuo AI free?

The Hailuo app gives daily free credits, more during launches. The API and hosts charge per clip or per second.

What is the difference between Hailuo 2.3 and H3?

2.3 is the late-2025 model at a flat price per six or ten seconds; H3 is the 2026 omni-modal model with native audio, 2K output, per-second pricing and open weights. Both pages run the same nine clips.