Kling v3 Pro audio on adds 50% per second: what Sume bills
fal lists Kling v3 Pro at $0.112 a second with audio off and $0.168 with audio on. Sume bills $0.14 and $0.21 a second for the same two settings.

Switching audio on for Kling v3 Pro raises the price by exactly 50 percent. fal's model page, read on 2026-10-04, lists $0.112 a second with audio off and $0.168 a second with audio on. The same page lists $0.196 a second when voice control is used with audio. Sume's estimator for the model uses the same two list rates and applies its 1.25 factor.
Vendor list versus Sume billed
The Sume column is the provider list price times 1.25, which the video docs state applies to every model. The 5-second figures are from the repo's Kling v3 Pro estimator, called once with audio generation off and once with it on.
| Setting | fal list per second | Sume billed per second | Sume 5 s clip |
|---|---|---|---|
| Audio off | $0.112 | $0.14 | $0.70 |
| Audio on | $0.168 | $0.21 | $1.05 |
When audio is worth the extra 7 cents a second
On Sume the difference is $0.07 a second. For a 5-second clip that is $0.35, for a 15-second clip $1.05. If you are going to lay a voiceover or music bed over the clip in an edit anyway, generate with audio off and keep the clip cheaper. If the clip's own sound is the point, such as ambient room tone or a spoken line, audio on avoids a second step.
A quick rule
- Audio you will replace: off.
- Audio you will keep: on, and price it at $0.21 a second, not $0.14.
- Voice control is a third tier on fal's page ($0.196 list). It is not priced here on Sume, so check the models endpoint before assuming a figure.
The same 50 percent at every length
Because both rates are per second and the ratio is fixed at 1.5, the audio premium scales with length. A 5-second clip costs $0.35 more with audio on, a 10-second clip $0.70 more, and a 15-second clip $1.05 more. Across a batch of 100 ten-second clips the audio switch alone is $70.00 on Sume.
That is the number to put in front of whoever approves the budget. If the sound would be replaced in the edit, the audio-on premium on a batch is money spent on something you delete.
Always confirm the live rate. The video docs point to GET /v1/videos/models for each model's prices, and list rates at the vendor can change.
Sources
Related posts
More in Pricing
- LTX-2.5 audio-to-video price vs Sume audio references
LTX-2.5 audio-to-video is billed per second of input audio: $0.13 Fast, $0.17 Pro. Sume has no audio-to-video id; audio references need an image or video too.
- Luma's ElevenLabs Music at 98 credits a minute vs ElevenLabs and Sume
Luma lists ElevenLabs Music v1 at 98 credits a minute: $0.16 to $0.29 by plan. ElevenLabs lists $0.15 a minute; Sume charges a flat $0.125 per generation.
- Luma ElevenLabs v3 voice: 21 credits per 1,000 characters in dollars
Luma lists ElevenLabs v3 TTS at 21 credits per 1,000 characters: $0.035 to $0.063 by plan. ElevenLabs lists $0.08 on its API and Sume lists $0.0475.
- H3 Max on Luma at 18, 28 and 56 credits a second, in dollars
Luma lists H3 Max at 18, 28 and 56 credits a second for 480p, 768p and 1080p: $0.03 to $0.17 by plan. fal lists $0.05; Sume's lip-sync is $0.0625 to $0.20.
Written by Sume