Wan 3.0 and MiniMax H3 Max clip price on Sume at 5, 10 and 15 seconds

At list times 1.25, a 10-second Wan 3.0 clip is $0.625 at 480p, $1.25 at 720p and $2.50 at 1080p. MiniMax H3 Max is $0.625, $1.00 and $2.00 at its three tiers.

5 min readSume
All posts

Sume bills every Video Router row at list times 1.25, so a 10-second Wan 3.0 clip is $0.625 at 480p, $1.250 at 720p and $2.500 at 1080p, before the cent rounding the docs mention. MiniMax H3 Max is $0.625, $1.00 and $2.00 for its three tiers. Read 2026-10-03 from the Video Router docs.

The docs say the billable amount is list times 1.25 with a ceil to cents, so I have shown three digits and left the rounding to the invoice.

The table

List rates come from the catalog table in the Video Router docs: Wan 3.0 at $0.05, $0.10 and $0.20 a second for 480p, 720p and 1080p, and H3 Max at $0.05, $0.08 and $0.16 for 480p, 768p and 1080p. Wan 3.0 runs 2 to 30 seconds per job and H3 Max 5 to 15 seconds, so the 5, 10 and 15 second columns all fit H3 Max, and Wan 3.0 fits all three too.

Sume price per clip at list x 1.25 (unrounded) (read 2026-10-03)
ModelResolution tierList per secondBilled per second5 seconds10 seconds15 seconds
Wan 3.0480p$0.05$0.0625$0.312$0.625$0.938
Wan 3.0720p$0.10$0.1250$0.625$1.250$1.875
Wan 3.01080p$0.20$0.2500$1.250$2.500$3.750
MiniMax H3 Max480p tier$0.05$0.0625$0.312$0.625$0.938
MiniMax H3 Max768p$0.08$0.1000$0.500$1.000$1.500
MiniMax H3 Max1080p$0.16$0.2000$1.000$2.000$3.000

What the ratios say

Wan 3.0 steps up in doubles: 720p costs twice 480p and 1080p costs twice 720p. H3 Max is gentler at the middle step (1.6 times) and doubles to the top. At 1080p, H3 Max at $0.20 a second is $0.05 a second cheaper than Wan 3.0 at $0.25.

For a rough draft loop, 480p on either model is the same $0.0625 a second. Three drafts of 5 seconds each cost $0.938, and the final render at 1080p of 10 seconds costs $2.00 on H3 Max or $2.50 on Wan 3.0.

Native versus refined 1080p on H3 Max

The docs state that H3 Max's 1080p is a latent refinement from 768p, with no 2K or 4K, and native resolutions are 480p and 768p. So the cheapest way to see what a prompt does is 480p, and the step to 1080p is a refinement pass, not a fresh render. The Wan 3.0 row lists audio and three input modes; H3 Max lists native stereo audio with no toggle.

Limits

Read the per-model capabilities from the catalog endpoint before you pin an id.

  • Prices are list at the time of the docs' fal reference dates; they change.
  • Cent rounding can add up to one cent per job.
  • Quality and motion differ between the models; price is not the only axis.

Sources

Related posts

More in Pricing

All Pricing posts

Written by Sume