Does a 60 fps Timeline render cost more than 24 fps on Sume?

No. The public rate for a Timeline 1.0 render is $0.10 per output minute, rounded up, and output.fps accepts only 24, 25, 30 or 60. The docs list no frame-rate term in the price, so a 60 fps render of 30 seconds costs the same $0.10 as a 24 fps one.

What the frame rate changes is smoothness. If the rate you pick differs from the native rate of a source, the render repeats or drops a frame every few frames and warns output_fps_resamples_sources. This page covers both halves, with data read from the Timeline 1.0 page on 2026-10-10.

What the price depends on

The billing unit is the output minute. The reserve is ceil(audio.duration_seconds / 60) minutes, and the audio spine length is the output length, with a range of 1 to 1800 seconds. The job calls no provider inference, only the worker ffmpeg, so there is no per-pixel or per-frame term in the docs.

That is also why the useful question is length, not rate. A 45-second ad at 60 fps and the same ad at 24 fps are both one billable minute, $0.10 each, while the same ad stretched to 61 seconds doubles the bill. Check the length you actually need before you tune the frame rate, because length is the only lever that moves the price.

The minute ladder

The cap is the one-line story: 1800 seconds is 30 minutes, and 30 x $0.10 = $3.00. The first second past a minute boundary costs a whole extra minute, so cutting 61 seconds down to 60 saves $0.10.

What the fps field does

If you omit output.fps, the job renders at the rate of the sources. The longest video sources decide it, stills have no rate, and 30 applies only when no source has a rate. That default is usually the best choice, because it never repeats or drops a frame.

Set it explicitly only when the destination demands it. Then conform your clips to that rate first. A video trim with output: { fps: 30 } conforms one clip in the same $0.02 job that cuts it, and the render then has nothing to resample.

What to do about judder

Mixed-model timelines are the usual cause. Clips from different generators often come out at different native rates, and a forced 30 fps render of 24 fps material shows a rhythm of repeated frames on pans. The warning output_fps_resamples_sources carries the rate the sources wanted, so read it from the result, set your output.fps to that value if the platform allows, and rerun.

A rerun is another render charge, so test with a short 10-second timeline first. At $0.10 per started minute it is cheap, and a plan call, which is unbilled, tells you the billable minutes before you spend anything.

Other limits that matter

Fades have their own limits: fade_in_seconds and fade_out_seconds run from 0 to 5 s, and their sum must be at most the output length. Transitions are fade, wipeleft, wiperight, slideup, slidedown or dissolve, last at most 1 second, and use at most 50 percent of the shorter neighbor. None of these change the price.

The default output size is 1080 by 1920. For a compose shot feeding the render, set the compose output to the same size so the clip is not rescaled twice; compose is a flat $0.02 job. The audio layer follows the spine: Timeline audio concat or split is $0.01 per job and joins up to 20 parts without silence at the seams, so a separate concat adds one $0.01 job, not another per-minute line.