A 17-second Kling 3.0 Motion Control clip: $2.68 on Sume

A 17-second Kling 3.0 Motion Control clip costs 17 x $0.1575 = $2.6775 on Sume, which rounds to $2.68. Motion Control is priced per output second, from 1 to 30 s, so the price is linear: 1 s is $0.1575 and 30 s is $4.725.

Price by length

From the Sume catalog facts for the kling-motion-control entry, read 2026-10-09. Kling 3.0 Motion Control is the only Kling model Sume lists. Sume does not list Kling 4.0, and its API price was not public when this page was written.

Budget checks

With $5 you can cover one 30-second run ($4.725) and have $0.275 left. Because the hold is the full price at submit, a balance below $2.6775 returns 402 insufficient_credits for the 17-second run, even if shorter runs would fit.

The usage page says a failed job, or one canceled before capture, releases its reserved amount, so a failed motion-control run does not leave a charge.

Motion Control next to the text-to-video rows

At $0.1575 per second, Motion Control sits between Wan 3.0 at 720p ($0.125) and Seedance 2.0 at 480p ($0.17577) in per-second terms. A 17-second Wan 3.0 clip at 720p costs 17 x $0.125 = $2.125, so the Motion Control run is $0.5525 more for the same length. They are not interchangeable: Motion Control transfers the movement of a reference onto your subject, while the text-to-video rows invent the movement.

If you also need audio under the clip, the lip sync and music prices are separate catalog entries, so price each step on its own line.

Because the price is linear, you can cost a whole batch with one multiplication. Twenty runs of 17 s are 20 x 17 x $0.1575 = $53.55, and the same twenty runs trimmed to 10 s are 20 x 10 x $0.1575 = $31.50. The saving, $22.05, comes entirely from the seven seconds you did not need.

Choosing the length

Match the output length to the reference motion rather than picking the maximum. If your reference move is 17 s, a 30-second output pays for 13 s you do not need: 13 x $0.1575 = $2.0475. Also keep in mind that Motion Control is not a text-to-video row, so compare it with the other Sume video rows only on price per second, not on what each one produces.