Kling 3 audio on adds 7 cents a second: 18 clips of 9 seconds on Sume

Audio costs 7 cents a second more on Kling Video v3 Pro on Sume, so eighteen 9-second clips cost $22.68 with audio off and $34.02 with audio on, a difference of $11.34. The video-router catalog lists list prices of $0.112 a second with audio off and $0.168 with audio on; Sume sells at list times 1.25, rounded up to the cent per job, which is $0.14 and $0.21 a second.

What the catalog lists for Kling 3

Kling Video v3 Pro (kling-3) accepts text-to-video, image-to-video and an end frame. It does not take reference images, videos or audio. Resolutions are 720p and 1080p, aspect ratios 16:9, 9:16 and 1:1, and durations 4 to 15 seconds. The price does not change between 720p and 1080p, only with the audio switch.

That makes the audio toggle the one real cost lever on this row. A silent clip you plan to score yourself costs a third less than one with generated sound.

The price table

Each job is rounded up to the cent, so the per-second figure drives the total but not to the cent. At 9 seconds the unrounded audio-off price is $1.26 exactly, and audio on is $1.89 exactly.

When audio is worth it

If the clips are b-roll under a voiceover or a music bed, audio off saves 7 cents a second and you add sound later. If the clip is a person speaking or an object making a sound that carries the shot, generated audio at $0.21 a second is the cheaper path against building it yourself.

A batch decision follows. For 18 clips at 9 seconds, the audio-on premium is $11.34. A caption job is $0.20, so for the same money you could burn captions onto about 56 clips.

Plan the job count too. A 9-second clip is a single job, so 18 clips are 18 jobs and the table total is simply 18 times the per-clip price. A 20-second spot needs two jobs at the 15-second maximum, which is worth pricing as 15 plus 5 seconds rather than 20 times the rate.

Because the price does not depend on resolution, run Kling 3 at 1080p whenever the audience will see it at that size. Only the audio switch changes the bill.

• Kling 3 has no reference inputs on Sume, so brand consistency comes from the first frame.
• The 15-second maximum means a 30-second video is two jobs and a timeline join.
• Compare with Wan 3.0 at 720p, which is $0.125 a second with audio listed as supported on its row.