MiniMax H3 vs H3 Max: a 5-second clip priced at each resolution

A 5-second MiniMax H3 clip costs $0.3125 at 480p and $0.375 at 768p on Sume. The same length on MiniMax H3 Max costs $0.3125 at 480p, $0.50 at 768p and $1.00 at 1080p. Both accept 5 to 15 seconds, and 5 is the minimum.

H3 Max is the more expensive one at 768p (about 1.33 times H3), and it is the only one of the two with a native 1080p tier. Figures are provider list x 1.25 as of 2026-10-08.

Per-second rates

H3 lists at $0.05 per second at 480p and $0.06 at 768p. H3 Max lists at $0.05, $0.08 and $0.16. Multiply by 1.25 for Sume: H3 is $0.0625 and $0.075 per second, and H3 Max is $0.0625, $0.10 and $0.20. Sume's test fixture reserves $0.375 for a default 5-second H3 clip and $0.50 for a default 5-second H3 Max clip, both at 768p.

The fal page for H3 Max describes it as a post-trained variant of MiniMax H3, tuned for stronger prompt adherence and better aesthetics, and states the post-launch rates of $0.05 per second at 480p and $0.08 per second at 768p (fal page, read 2026-10-08). The 1080p rate comes from the Sume catalog.

What 1080p means on H3 Max

The Sume catalog describes H3 Max 1080p as a latent refinement from native 768p. Native output is 480p or 768p, and Sume lists no 2K or 4K tier for H3 Max. MiniMax H3 is different: Sume bills 2K and 4K upscales if your request includes them, at list $0.13 and $0.16 per second, on top of its native 480p and 768p.

Which to choose

For a 15-second clip at 768p, H3 costs 15 x $0.075 = $1.125 and H3 Max costs 15 x $0.10 = $1.50. The $0.375 difference buys the post-trained variant and the option to go to 1080p at 15 x $0.20 = $3.00. If you plan to send more than five reference images to H3, add $0.10 for each one beyond the fifth before you compare the two.

Shared envelope

Pick H3 for cheap drafts and long reference sets; pick H3 Max when prompt adherence or a 1080p deliverable matters.

• Duration 5 to 15 seconds, aspect ratios adaptive, 21:9, 16:9, 4:3, 1:1, 3:4 and 9:16.
• Text-to-video, first and last frame image-to-video, and reference-to-video, with native stereo audio and no audio toggle.
• Reference images up to 9, videos up to 3 and audio up to 3, with a total of 12 or fewer; audio cannot be the only reference.
• H3 charges list $0.08 for each reference image after the first five; H3 Max bills output seconds only.