Six Motion Control jobs: 29.8 s of footage bills 32 s, $5.04
Six Kling 3.0 Motion Control jobs of 4.5, 6.2, 3.8, 5, 7.4 and 2.9 seconds bill 5+7+4+5+8+3 = 32 seconds on Sume: 32 x $0.1575 = $5.04.

Six Kling 3.0 Motion Control jobs on Sume with driving clips of 4.5, 6.2, 3.8, 5.0, 7.4 and 2.9 seconds bill 5 + 7 + 4 + 5 + 8 + 3 = 32 seconds in total. At $0.1575 per output second that is $5.04, even though the footage adds up to only 29.8 seconds.
Rounding happens per job
The catalog says the route reserves ceil(duration_seconds) x the Kling list price after Sume's margin, with output length following the driving video. Each job rounds on its own, so six jobs round six times.
Two practical notes. The output length follows the driving video, so the 'Bow' job is a 3-second clip, and a short move is cheap. And because each job is independent, a failed job does not affect the other five; the catalog says failed jobs are refunded before capture, so only the successful clips stay on the bill.
| Move | Driving clip | Billed seconds | Cost |
|---|---|---|---|
| Wave | 4.5 s | 5 | $0.7875 |
| Jump | 6.2 s | 7 | $1.1025 |
| Point | 3.8 s | 4 | $0.63 |
| Spin | 5.0 s | 5 | $0.7875 |
| Dance | 7.4 s | 8 | $1.26 |
| Bow | 2.9 s | 3 | $0.4725 |
| Total | 29.8 s | 32 | $5.04 |
The cost of rounding
One 29.8-second driving clip would bill 30 seconds, $4.725, so doing six moves as six jobs costs $5.04 - $4.725 = $0.315 more. That is the price of getting six separate clips, each a different move, which you can arrange and reuse. A single long clip would give you one continuous motion.
If you can shape your reference footage to land on whole seconds, you save up to the difference: 29.8 s of raw footage would be 29.8 x $0.1575 = $4.6935, so the rounding overhead is $0.3465 in total, about 7 percent of the bill.
If you plan to reuse the moves, store the six result clips with the move name and the driving-clip length. That turns a one-off $5.04 spend into a reusable library for the mascot, and later jobs can reuse the same driving clips only if you want the same motion again.
- Use video trim ($0.02 flat) only when a cut crosses a whole-second boundary.
- Keep the same mascot still for all six jobs so the character stays consistent.
- Each job needs its own Idempotency-Key.
Scaling the set
Repeat the set for four mascots and the billed seconds are 4 x 32 = 128, or $20.16. The maximum of any single job is 30 seconds ($4.725), so no single driving clip can exceed that.
Document the driving clips themselves. A 'Spin' clip that runs 5.0 seconds exactly costs 5 x $0.1575 = $0.7875, but the same spin filmed at 5.1 seconds costs 6 x $0.1575 = $0.945. If you film the references yourself, a countdown and a clapperboard moment let you cut to the second in an editor before upload. That small habit removes most of the $0.3465 of rounding overhead, and the clips become reusable assets in your library.
Sources
Related posts
More in Media tools
- Does Kling motion control follow my prompt? Only for appearance
On Sume's Kling 3.0 Motion Control the optional prompt (up to 2,000 characters) steers appearance only; motion and framing follow the driving video.
- Korean captions: slam returns 400, so use korean-ad with language ko
Latin caption styles have no Hangul glyphs, so Korean text on slam, punch or tiktok-green returns 400 caption_hangul_text_latin_style. Use korean-ad.
- LinkedIn 1200x675 video ad: Sume needs even sizes, so use 1280x720
LinkedIn lists 1200x675 as a 16:9 size, but 675 is odd. Sume timeline sizes must be even, so render 1280x720 or 1920x1080, both inside LinkedIn's range.
- LinkedIn 30-minute video: 24 spot-check stills, one every 75 seconds
LinkedIn allows videos up to 30 minutes. Video inspect takes 1,800 seconds and 24 stills per call, so at[] with 37.5 + 75k covers the whole file in one job.
Written by Sume