3-minute Reels on Timeline: 61 seconds bills as two minutes

Sume Timeline bills $0.10 for each started minute of output, so the length you declare matters in whole-minute steps: a 60-second Reel is $0.10, a 61-second one is $0.20, and a full 180-second Reel is $0.30 (read 2026-10-03). If your cut lands a second or two over a minute boundary, trimming the tail is the cheapest edit you can make.

The reason to care is the format. Metricool's Instagram news page reports that Instagram's updated Reels guide gives ideal clip lengths from 3 seconds to 3 minutes (reported, read 2026-10-03). Longer cuts mean a pipeline that renders each revision pays for every pass.

How the bill is computed

The Timeline docs state the rate as $0.10 per ceil of the output minute, and the reserve is ceil(audio.duration_seconds / 60) minutes. audio.duration_seconds is required and runs from 1 to 1,800 seconds, so it is also your output length. The job runs on the worker's ffmpeg with no provider inference, which is why the price is flat per minute rather than per model.

Two consequences follow. First, the price is set by the length you declare, not by the number of clips or transitions you place. A 12-slot cut and a 40-slot cut of the same length cost the same. Second, it is set by the declared spine, so a cut that you plan at 58 seconds but stretch to 62 with a late end card has moved to the higher bracket.

A length check before you submit

Timeline has a /plan route that is unbilled, and a plan is the right place to catch a bracket crossing before you pay. For a quick guard in your own code, compute the billed minutes from the declared length and warn when the cut is only a few seconds past a boundary.

The helper below prints the bracket and how many seconds you could shave to drop a minute. It does not call Sume, so it is safe to run on every revision.

import math

def bracket(seconds: float) -> dict:
    minutes = math.ceil(seconds / 60)
    over = seconds - (minutes - 1) * 60
    return {"seconds": seconds, "billed_minutes": minutes,
            "cost_usd": round(minutes * 0.10, 2),
            "trim_to_save": round(over, 1) if minutes > 1 and over <= 5 else 0}

for s in (58, 61, 90, 121, 179.5):
    print(bracket(s))
# 61 s -> 2 minutes, $0.20, shave 1.0 s to fall back to $0.10

What to do with a long Reel

Decide the length from the story and then round down to a boundary only if the last seconds add nothing. A closing beat is worth ten cents. A second of dead air at the end is not.

Revise in a cheap loop. Timeline plans cost nothing, so check structure, slot starts and warnings with /plan first and render only when the plan is clean. Each render is a full-length charge, so a ten-pass revision of a 150-second Reel is $3.00 in Timeline alone, against $0.30 for a single final render.

Put the fade where you want it. output.fade_out_seconds is 0 to 5 seconds and counts inside the length you declared, so a three-second fade does not extend the file past your bracket. Check that the fade-in and fade-out sum is no longer than the output.

Use one declared length per episode and store it with the episode record. A season of eight 150-second episodes is eight renders at $0.30, which is $2.40. If one episode drifts to 181 seconds, the bracket jumps to four minutes and that render is $0.40, a small number that is easy to miss and worth a guard in the submit script.

Finally, remember that the clips are the real cost. Timeline is ten cents a minute; video generation is billed per clip by model. Use this table to avoid waste at the end of the pipeline, and read the live catalog for the generation prices before you plan the middle.

What Sume does and does not do

Sume renders the cut you declare, up to 1,800 seconds, and bills started minutes. It does not shorten your video to fit a bracket, and it does not know Instagram's upload rules; check the length limit in the app you post from. The Metricool figures above are a third-party report of the feature, not an Instagram specification.