Music bed for a 3-minute video: the render and track cost $0.425

Adding a music bed to a three-minute video on Sume costs $0.425 at list rates before the clips themselves: $0.30 for the Timeline render and $0.125 for one generated track. The pricing page, read 2026-10-03, lists Timeline render at $0.10 per output minute and Music at $0.125 per audio. The Music 1.0 docs say the track is steered by the prompt, for example by asking for a track of a given length, since there is no duration field.

What goes into the $0.425?

Clips, narration and captions are priced on their own lines, so this is only the layer that carries sound.

How do I get a track long enough?

Write the length into the prompt, such as a three-minute track, and use section markers like [0:00-0:30] Intro: to shape it. Music 1.0 is described as full-length structured songs up to a few minutes, so three minutes is within the described range, but nothing promises exact duration. If a track comes up short, join two takes with Timeline audio concat, which joins in the sample domain with no gap at the seams, and check the seam by ear.

What is not in the price?

• Rejected or retried music generations are separate $0.125 lines each, because every accepted generation is charged.
• The Timeline render price is per output minute; I did not verify how partial minutes are rounded.
• Whether the music fits the picture is a judgement call. Audition several takes if the mood matters, as the five-takes post in this series explains.