Music for 21 scenes: 1 to 5 takes each at $0.125 flat on Sume

Scoring a 21-scene video on Sume Music 1.0 costs $2.625 with one track per scene, $5.25 with two takes, $7.875 with three and $13.125 with five, because the price is a flat $0.125 per generation regardless of prompt. The hold on your balance is slightly higher, since each job reserves 13 cents.

Flat pricing changes the planning

The catalog says Sume Music 1.0 charges a fixed $0.125 per audio generation and that the price does not vary by prompt or image input. A short sting and a long bed cost the same, so the only variables are how many generations you run and how many you keep.

That means retakes are linear: three takes for each of 21 scenes is 63 generations, 63 x $0.125 = $7.875. If you keep one of three, each kept track costs $0.375.

List cost versus the hold

The catalog's estimate for one generation is 13 cents, which is $0.125 rounded up to a whole cent for the balance check. For 63 jobs the hold is 63 x $0.13 = $8.19, about 31.5 cents above the list total of $7.875. That is a temporary hold while jobs run, and the catalog says completed jobs are captured and failed ones refunded.

If your balance is close to zero, size it against the hold, not the list price. With only $8.00, 63 jobs at 13 cents would not all fit at once even though the list total is $7.875.

Choosing the number of takes

One take is fine for background beds where any plausible track works. Use three takes for scenes where mood matters, and five only for the hero scene. A mixed plan: 15 scenes x 1 take, 5 scenes x 3 takes, 1 hero scene x 5 takes is 15 + 15 + 5 = 35 generations, or $4.375.

Compare against the stills: eight Nano Banana 2.1 images at 1K are 8 x $0.10 = $0.80, so music for 21 scenes can cost more than the visuals in a still-heavy edit.

• Mixed plan: 35 generations = $4.375
• Three takes everywhere: 63 generations = $7.875

A second scenario

Compare music with the visuals. A 21-scene edit with an 8-second Wan 3.0 clip per scene at 480p costs 21 x 8 x $0.0625 = $10.50 for the visuals. Three takes of music at $7.875 is 75 percent of that, so music retakes are a real part of the budget. One take per scene is $2.625, a quarter of the visuals.

Where it goes next

Generated tracks can be placed in a Timeline render, which is priced separately at $0.10 per output minute. The Music and Timeline docs pages describe the inputs.

Reproduce the numbers

Open https://api.sume.com/v1/catalog, which needs no key, and find the entry for the product. Each model_pricing row carries price, unit, pricing_basis and the estimated, minimum and maximum cents. Multiply the unit price by your quantity and round only at the end. The catalog also has applied_* fields that include an agent fee of 550 basis points; the figures in this post use the base prices.