Timeline bills 61 s as two minutes: the 60 s cut that halves the fee

Timeline 1.0 charges $0.10 per output minute rounded up, so a 60-second render costs $0.10 and a 61-second render costs $0.20 (as of 2026-10-08). The fee is small, but it is easy to plan around when your shots are 30 seconds long.

Where the steps fall

The rule is ceil(audio.duration_seconds / 60) x $0.10. The steps are at 60, 120, 180 s and so on.

Two shots with a fade

Two 30 s shots with a 0.5 s crossfade overlap by 0.5 s, so the picture would run 59.5 s. The declared audio length is what bills, and the docs say video coverage may end at most 0.5 s before the spine ends. If the spine is 60 s, you pay for one minute.

If you declare 61 s so the last frame holds, you pay for two. Keep the spine at 60 s unless the story needs the extra second.

Is the fee ever worth optimising?

Against generation costs, rarely. For two 30 s shots on Wan 3.0 at 720p the generation is $7.50, so $0.10 is 1.3% and $0.20 is 2.7%. On Seedance 2.5 at 720p ($34.68) it is 0.3% and 0.6%.

It matters when you render many short programs: a thousand 61 s edits cost $200 in render fees against $100 for the same edits at 60 s. Check the plan endpoint, which returns billable_minutes before you create the job.

Planning checklist

Count shot seconds before generation: 2 x 30 s fits one minute, 2 x 30 s plus a 5 s end card needs two. Cutting the end card, or generating shots of 28 s each, keeps you at 60 s. The grid prices 25 and 30 s but not 28, so confirm a 28 s price first.

Declare the audio spine at the exact length you want. The plan endpoint is free and returns billable_minutes, so run it as part of every build.

Check the number yourself

To confirm the figure behind this post, read the seedance-2.5 entry from GET /v1/video-router/models and check that 30 s and 720p are inside its duration_seconds and resolutions. The grid price for that cell is 1734 cents ($17.34). It is the provider list price times 1.25, rounded up to the next cent for the job, as of 2026-10-08.

Use an Idempotency-Key on every create call so a retry cannot start a second paid job, then poll the job until it is complete and download the file from the result. Media inputs for Timeline or for references must be hosted files, so import them first.

None of the prices here say anything about picture quality, motion or how well a model follows a prompt. Prices can change when a provider changes its list, which is why every table carries a date. Re-read the catalog before you commit a large batch, and run one test job at the exact setting you plan to use.