Timeline render: 61 seconds costs double 60 on Sume
Sume Timeline reserves $0.10 per output minute, rounded up: a 60-second render is $0.10, a 61-second render is $0.20, a 30-minute render is $3.00.

A 60-second Timeline render on Sume reserves $0.10 and a 61-second render reserves $0.20, because the reservation is ceil(seconds / 60) x $0.10, with a one-minute minimum and a 30-minute maximum (1,800 seconds). Trim a render by one second and you can halve its reservation.
The rule
The catalog's pricing basis for Timeline 1.0 is ceil(audio.duration_seconds / 60) x $0.10 per output minute, maximum 1,800 seconds. The same text says the render uses worker ffmpeg only, with no provider inference, and that all source URLs must be Sume-hosted.
Every length from 1 to 60 seconds reserves $0.10, from 61 to 120 reserves $0.20, and so on. The table shows the edges: 29 s and 60 s are both one minute; 601 s becomes 11 minutes.
| Output seconds | Billed minutes (ceil, min 1) | Reservation |
|---|---|---|
| 29 | 1 | $0.10 |
| 60 | 1 | $0.10 |
| 61 | 2 | $0.20 |
| 600 | 10 | $1.00 |
| 601 | 11 | $1.10 |
| 1799 | 30 | $3.00 |
| 1800 | 30 | $3.00 |
Reservation is a ceiling
The catalog description adds that the render is reserved at the per-minute rate and captured at the render's own compute cost, never above the reservation. So $0.20 for a 61-second render is the most you can be charged, not a guaranteed charge.
That makes the reservation the right number for budgets and the wrong number for invoices. Plan with the ceiling; expect the captured amount to be at or below it.
Where it matters
For one render, the difference between 60 and 61 seconds is ten cents. For a library it adds up. A hundred 61-second renders reserve 100 x $0.20 = $20.00; trimming each to 60 seconds reserves $10.00. If you can lose one second of padding per video, you save half.
Check durations in code before you submit: if the total is between 60 and 65 seconds, see whether trimming the outro by a few frames is acceptable. Do the same at 120, 180 and 600 seconds.
- 60 s: $0.10
- 61 s: $0.20
- 120 s: $0.20
- 121 s: $0.30
A second scenario
Check the other edge of the scale. A 25-minute render (1,500 seconds) reserves 25 x $0.10 = $2.50; a 1,501-second render reserves 26 x $0.10 = $2.60. The jump is the same ten cents, but at 30 minutes the cap stops it: 1,800 seconds is $3.00 and anything longer is outside the model's range, so split longer programs into parts.
Related per-job prices
Other Timeline operations are cheap flat jobs in the catalog: compose is $0.02 per job, audio concat or split is $0.01 per job, and video trim and filter are $0.02 per job. The Timeline docs page lists the slots and limits, including 1 to 200 video slots per render.
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.
Sources
Related posts
More in Pricing
- 10 hours of audio: Sume STT $6.00 vs Scribe v2 $2.20
Transcribing 10 hours of customer calls: Sume's $0.01 per audio minute is $6.00; ElevenLabs lists Scribe v2 at $0.22 per hour, $2.20. Read 2026-10-09.
- 15 hours of audio: Sume STT $9.00 vs Scribe v2 $3.30
Transcribing 20 podcast episodes of 45 minutes: Sume's $0.01 per audio minute is $9.00; ElevenLabs lists Scribe v2 at $0.22 per hour, $3.30. Read 2026-10-09.
- 25 hours of audio: Sume STT $15.00 vs Scribe v2 $5.50
Transcribing 25 hours of lecture recordings: Sume's $0.01 per audio minute is $15.00; ElevenLabs lists Scribe v2 at $0.22 per hour, $5.50. Read 2026-10-09.
- 37 hours of calls on Sume STT: $22.20 in 222 ten-minute chunks
Sume STT 1.0 is $0.01 per audio minute and one request reserves at most 10 minutes ($0.10). 37 hours is 2,220 minutes, so 222 chunks and $22.20.
Written by Sume