Music for 63 videos: 9 briefs, 3 takes each, $3.375 at 12.5 cents
Music 1.0 is a flat $0.125 per generation, with no duration parameter. 63 videos on 9 shared briefs cost $1.125; three takes per brief cost $3.375.

Music 1.0 on Sume costs a flat $0.125 per generation, and the model has no duration parameter, so length is steered in the prompt. That makes the number of generations, not the number of videos, the cost driver. Sixty-three videos that share 9 briefs need 9 tracks, or 9 x $0.125 = $1.125; if you generate 3 takes per brief and keep one, it is 27 generations and $3.375.
Three ways to cover 63 videos
Music 1.0 runs on Google Lyria 3.5 and makes structured songs of up to a few minutes, per the Music 1.0 doc. Price below is the flat generation fee only.
| Strategy | Generations | Math | Cost |
|---|---|---|---|
| One track per video | 63 | 63 x 0.125 | $7.875 |
| One track per brief, 9 briefs | 9 | 9 x 0.125 | $1.125 |
| Three takes per brief, keep one | 27 | 27 x 0.125 | $3.375 |
| One track per brief, plus 20 percent redo | 11 | ceil(9 x 1.2) x 0.125 | $1.375 |
The cost is small; the choice is creative
The gap between the cheapest and most expensive strategy is $6.75. Compared with video footage, where one 8-second Seedance 2.5 clip at 720p is $4.6224, a full music library for all 63 videos is under two clips.
So spend the difference on takes. A brief that names tempo, key and one named moment (the doc lists seven axes) produces a track you can reuse across a series, and three takes of the same brief give you a choice at $0.25 extra per brief.
Per-video cost and retention
Nine tracks across 63 videos means seven videos per track, so the music line is $0.125 / 7 = $0.0179 per video. With three takes per brief, it is $0.375 / 7 = $0.0536 per video. Neither number moves a video budget.
What moves the outcome is retention. Store each chosen track's URL against its brief and reuse it, because a second generation of the same brief is a new $0.125 charge and a different song. Keep the three rejected takes for a month in case the client changes the brief; they cost $0.25 per brief and nothing to store.
Music 1.0 also accepts an optional image for conditioning, which does not change the flat price in the catalog. Test one conditioned track against one prompt-only track on the same brief before you adopt either for all nine.
Where the cost does matter
Two cases change the picture: a pipeline that generates music per video in a loop with no cache, and an agent that retries on a vague quality judgment. Cap the number of generations per brief in code, store the track URL against the brief, and the line stays rounding error.
A cap you can enforce
Put the music budget in code so that no loop can spend more than planned.
- Cap generations per brief at 3, which is $0.375 per brief.
- Cap total music generations per campaign at 27, which is $3.375 for the nine-brief plan.
- Refuse a new generation if a track URL is already stored for that brief.
- Record the prompt next to the URL; the prompt is the only way to remake a track you liked.
Sources
More in Pricing
- 63,500-character chapter: Sume TTS $3.02 vs Aura-2 and ElevenLabs
A 63,500-character chapter costs $3.02 on Sume TTS in four requests. Aura-2 is $1.91, ElevenLabs v3 $5.08, and v4 Turbo $2.54 regular or $0.70 on promo.
- 640 Nano Banana 2.1 images at 2K: Google $16.13 batch, Sume $96.00
At 2K, 640 images cost $16.128 on Google Batch, $32.256 on Google Standard and $96.00 on Sume. Sume's Image API docs describe no batch discount.
- 660 GPT Image 2.5 images at three quality tiers: $27.95 on Sume
500 low, 120 medium and 40 high GPT Image 2.5 images cost $27.95 on Sume. Low 1K is $0.02475, medium 2K $0.055625, high 4K $0.2225 per image.
- $7.30 image wallet: how many images each Sume row buys
A $7.30 balance buys 294 GPT Image 2.5 low images, 166 Seedream 5 Lite, 73 Nano Banana 2.1 at 1K, 32 GPT Image 2.5 high at 4K, or 324 cutouts.
Written by Sume