MiniMax's H3 video model, generating 2K clips with native stereo sound from a text prompt. it is the lower-priced of the two H3 entries in the catalog, below H3 Max.
pick it if
you want 2K text-to-video with native stereo sound
you need clips up to 15 seconds
you want H3 at a lower per-second rate than H3 Max
skip it if
you need image or video references: this lane takes a text prompt only
a cheaper video model such as Wan 3.0 covers drafts
H3 Max is fal Research's post-trained version of the open-weights MiniMax H3 video model, tuned for prompt adherence and visual quality and served on a speed-optimised stack. it costs more per second than MiniMax H3.
pick it if
you want H3-based video with tighter prompt adherence and aesthetics
render speed matters: fal reports a 5 s clip in about 3 s
you need wide aspect ratios from 21:9 down to 9:16
skip it if
base MiniMax H3 at a lower per-second price is good enough
you need native 2K output; fal's 1080p is refined from 768p
faq
which is cheaper, MiniMax H3 or MiniMax H3 Max?
MiniMax H3 is cheaper — $0.2 / second vs $0.25 / second.
how do i use MiniMax H3 or MiniMax H3 Max?
in octomind cloud on a paid plan, ask the agent to generate with either model by name; it runs the job and returns the file.
try either model
switch between models inside the same octomind session.