Kling 3 sound on or off: a 12-second clip is $2.52 or $1.68 on Sume

On Sume a Kling 3 clip costs $0.21 per second with generated audio and $0.14 per second when you turn audio off. A 12-second clip is therefore $2.52 with sound and $1.68 silent, a $0.84 difference. The rule is simple: sound adds 50% to the per-second rate, so decide before you submit whether you need native sound at all.

Kling 3 in the video catalog is the Pro tier on fal. The repo's pricing basis is $0.112 per second with audio off and $0.168 per second with audio on at provider list, multiplied by Sume's 1.25. The row accepts 720p and 1080p, 4 to 15 seconds, and it does not accept reference_*_urls.

Price by length

The arithmetic

With audio: $0.168 x 1.25 = $0.21 per second, and 12 x $0.21 = $2.52. Silent: $0.112 x 1.25 = $0.14 per second, and 12 x $0.14 = $1.68. Dividing the two rates, $0.21 / $0.14 = 1.5. The fee column multiplies the with-audio price by 1.055: $2.52 x 1.055 = $2.66.

A common plan is to render silent and lay a TTS voiceover and a music track over it. For a 12-second clip that costs $1.68 + $0.0285 (a 600-character voiceover at $0.0475 per 1,000) + $0.125 (a Music 1.0 track) = $1.83, which is $0.69 less than Kling's built-in audio at $2.52. If the clip needs ambient sound that matches the action, native audio may be worth the premium.

Steps

• Set generate_audio: false on POST /v1/videos for the silent rate; leave it on for sound.
• Choose 720p or 1080p; the row lists both, and check the catalog for whether resolution changes the per-second price before you assume it.
• Keep durations at 4 to 15 seconds.
• Compare silent plus TTS plus music against native audio on one sample before you commit a series.

A 20-clip series

Twenty clips of 8 seconds each cost 20 x $1.68 = $33.60 with audio and 20 x $1.12 = $22.40 silent, so the audio premium for the series is $11.20. With the 5.5% platform fee from Sume's pricing policy those become $35.45 and $23.63.

At the longest length, 15 seconds, one clip is $3.15 with sound and $2.10 silent. If you plan to caption and voice every clip anyway, the silent route removes a cost you would otherwise pay twice.

What Sume does not do

Sume does not apply a separate fee for turning audio on; the audio price is part of the per-second rate. It does not accept reference URLs on Kling 3, so use Seedance, Wan or H3 Max for reference-driven shots. Rates were read from the repo on 2026-10-08 and fal states them as subject to change.