Three 61-second renders cost $0.60; one 183-second render costs $0.40

Three separate 61-second Timeline renders cost $0.60, because each one rounds up to 2 minutes at $0.10 per ceil(output minute). One 183-second render of the same footage costs $0.40, because ceil(183 / 60) is 4 minutes. Rounding happens once per render, so many short renders pay the rounding many times.

The arithmetic

The Sume docs state the public rate as $0.10 per ceil(output minute) and the reserve as ceil(audio.duration_seconds / 60) minutes. The table compares four ways to render roughly three minutes of material.

The cheap cases are exact minutes

The last two rows show the pattern. If every render lands on a whole minute, splitting costs the same as combining. The loss comes from the extra second. A 61-second render pays for a second minute to hold one second. Three of them pay that waste three times, 3 x 59 = 177 seconds of paid but unused time, while the combined 183-second render wastes 57 seconds.

So before you split a long edit into several files, check where each part ends. If you can make each part 60 seconds or less, the cost is the same as one combined render of the same total. If a part must be 61 seconds or more, merging it with a neighbour may be cheaper.

Do not merge past a platform limit

The cheaper plan is not always the right one. YouTube's help page for Shorts says a Short can be up to three minutes and must be a square or vertical video. A 183-second render is three seconds over that limit, so it would not qualify as a Short on that rule. For a Short, the target is 180 seconds or less, which is $0.30 as one render.

For other destinations, the Timeline limits are 1800 seconds of output and 200 video[] slots per render, and the render strategy chunks above 12 segments by default. These limits come from the Timeline docs.

• Combine parts when the platform allows the combined length and the story is one piece.
• Split when each part is its own deliverable, but aim at 60 seconds or less each.
• Use the unbilled POST /v1/timeline-1.0/plan to confirm billable_minutes before you submit.

Many short pieces versus one long one

The same logic applies to a reel series. Ten renders of 45 seconds each bill 10 minutes, $1.00, since each one rounds up to one minute. The same 450 seconds as a single render bills 8 minutes, $0.80. If the pieces do not have to be separate files, one render is cheaper. If they do, keep every piece at 60 seconds or less and the rounding costs you only the unused seconds inside each minute.

A quick check

Run this against your own list of planned lengths. It prints the cost of each file alone and the cost if you merged them, and it makes the rounding visible:

import math

def cost(seconds: int) -> float:
    return math.ceil(seconds / 60) * 0.10

parts = [61, 61, 61]
print("split   :", round(sum(cost(s) for s in parts), 2))
print("combined:", round(cost(sum(parts)), 2))