Can I render a Shorts season on Sume's Free plan?

Yes, in waves: the Free plan runs 1 job at a time and accepts 6, then returns 429 queue_full. Ten 30-second episodes need two submit waves.

4 min readSume
All posts

Yes, but in waves. Sume's Free plan processes one paid generation job at a time and accepts at most 6 (a queue of 5 plus the running one); the next submit fails with 429 queue_full.

So a season of ten 30-second episodes, one job each, is two waves of at most six jobs, with a pause for the first wave to drain.

The numbers

Generation concurrency is plan-only and prepaid top-ups do not raise it. The default queue capacity is the larger of 3 and five times the concurrency limit. Concurrency is a dispatch limit, not a submit limit: extra valid jobs queue as long as capacity remains.

Counting jobs, not episodes, is what matters: a 60-second Wan episode is at least two jobs.

Plan capacity from the Sume docs (read 2026-10-07)
PlanProcessingQueueAcceptedEpisodes of 2 jobs
Free1563
Pro4202412
Startup8404824
Scale2010012060

A wave plan

Submit up to six jobs, poll until some are terminal, then submit the next jobs. Use an idempotency key so a retry does not double-submit, and back off on 429. Balance is checked at submit: a shortfall is 402 insufficient_credits before provider work starts.

If you queue Format runs instead, the bulk route accepts 100 items but workspace concurrency still applies to the children; see the bulk post and the queue_full post.

  • Count jobs: clips plus renders plus captions.
  • Do the pilot at 480p to keep the pilot cheap.
  • Move up a plan when waves become the bottleneck.

Pilot first, then the season

On the Free plan the cheapest honest plan is a pilot. Use 480p at $0.0625 per second of Wan 3.0, which makes a 30-second beat $1.88. Six accepted jobs at that tier cost $11.25 at most. Run them, look at the result, and only then decide about a larger plan.

Waves have a rhythm. Submit six, wait for the first to complete and then top the queue back up to six. This keeps the running job busy and never reaches the queue_full error. Because Free processes one job at a time, a six-job wave is six sequential renders, so schedule it when you are not waiting for it.

If the waves become the bottleneck, the table above shows the next step: Pro accepts 24 jobs and runs 4 at a time. Concurrency is plan-only, so a prepaid top-up does not raise it. Read the errors post for what to do when a submit is refused.

Plan the calendar too. A Free-plan wave of six 30-second Wan beats at 480p is six sequential renders, so run it overnight or while you write the next batch of prompts. When a wave finishes, record what it cost from the receipts, and use that real number, not the table, to size the next wave.

Sources

Related posts

More in Pricing

All Pricing posts

Written by Sume