Kling 3 15-second clip: $2.10 silent, $3.15 with audio on Sume

A 15-second kling-3 clip costs $2.10 on Sume with generate_audio: false and $3.15 with audio on, which is the default. The wallet must hold at least that amount when you submit, or the API answers 402 insufficient_credits.

The rate is the provider list price times 1.25. Sume's pricing constants list Kling v3 Pro at $0.112 a second without audio and $0.168 with it. Multiply by 1.25 and you get $0.14 and $0.21.

The arithmetic

15 x $0.14 = $2.10. 15 x $0.21 = $3.15. The audio flag adds $0.07 a second, or $1.05 over 15 seconds, a 50% premium. Because generate_audio defaults to true, a request that never mentions it is billed at the higher rate.

How the 402 works

Sume reserves the cost at submit, then captures the final amount when the job completes. If the wallet balance is below the reserve, the API returns 402 insufficient_credits. A request with audio on at 15 seconds needs $3.15 free, even if you would have paid $2.10 with audio off.

Send an Idempotency-Key header on every submit so a retry after a timeout does not reserve and bill a second job.

Send it silent on purpose

If you add a voice-over or music in post, turn the audio off and pay $2.10.

curl -X POST https://api.sume.com/v1/videos \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: kling3-15s-silent-001" \
  -d '{"model":"kling-3","prompt":"A paper boat drifts down a rainy street gutter, handheld camera","duration":15,"resolution":"1080p","aspect_ratio":"16:9","generate_audio":false}'

Is the audio worth $1.05

For a 15-second clip, native audio costs $1.05 more. A separate voice-over or music track, added on a timeline, can cost less and gives you control over wording and timing. A talking face is another story, since a video model does not lip-sync to a later voice-over. The audio premium is the price of dialogue and sound produced together.

See the longer discussion in the audio surcharge post.

Budget a batch

Multiply before you queue. Twenty clips at 15 seconds with audio on is 20 x $3.15 = $63.00. The same twenty silent is 20 x $2.10 = $42.00, a difference of $21.00.

• Top up the wallet to at least the full reserve of the largest single job you will submit.
• Do not queue parallel jobs that together exceed the balance. Each reserve is held at submit.
• Log usage.cost from the finished job to compare the real charge with your estimate.
• Decide on audio per project, not per clip, so a series keeps one sound character.