30-second AI video with sound: what it costs on Sume

A 30-second video with sound costs $6.30 on Kling 3 as two 15-second 1080p jobs with audio on, $3.75 on Wan 3.0 at 720p and $17.34 on Seedance 2.5 at 720p. Kling 3 silent costs about two thirds of that: $4.20.

Audio switches

Kling 3 prices audio on and off separately: $0.21 a second with audio and $0.14 without, on Sume. The Video Router lists Wan 3.0 as an audio model; check generate_audio in the catalog row (GET /v1/videos/models) for the model you pick before you count on sound.

Reading it

Kling 3 is the only row here that forces 1080p, so its comparison point is Wan 3.0 at 1080p ($7.50). The 30-second Kling route is two jobs, and the seam at 15 seconds is yours to hide.

If sound is a voiceover rather than model audio, a separate TTS step is cheaper and gives you edit control; keep that as a separate line in the budget.