A full queue on each Sume plan: reserved dollars for 10-second jobs
If every accepted job holds its estimate, a full Scale queue of 120 ten-second Seedance 2.5 jobs at 1080p reserves $1,705.86. Free, Pro and Startup too.

A Scale workspace that fills its 120 accepted-job slots with 10-second Seedance 2.5 jobs at 1080p holds $1,705.86 in reservations; a Free workspace with 6 slots holds $85.29. Sume reserves the estimated amount when it accepts a job, and the accepted job capacity is the documented cap on processing plus queued jobs. Multiply the two to see the most a plan can have outstanding.
The arithmetic
Accepted job capacity per plan comes from the generation admission doc: concurrency plus queue capacity. A 10-second Seedance 2.5 job is 10 x $0.5778 = $5.778 at 720p and 10 x $1.42155 = $14.2155 at 1080p.
| Plan | Accepted jobs | At 720p | At 1080p |
|---|---|---|---|
| Free | 6 | $34.67 | $85.29 |
| Pro | 24 | $138.67 | $341.17 |
| Startup | 48 | $277.34 | $682.34 |
| Scale | 120 | $693.36 | $1,705.86 |
What this means for a balance
A balance smaller than the figure does not block submits until one reservation no longer fits; that submit fails with 402 insufficient_credits. On a Pro workspace with $200.00, 14 ten-second 1080p jobs reserve $199.02 (14 x 14.2155); the fifteenth submit fails and is not queued.
Reservations return on failure and on cancellation before capture, and completed jobs capture the amount. So the figure is a ceiling on what is held at once, not what you will spend.
Half-full queues and drafts
Queues are rarely full. A Pro workspace with 12 jobs accepted at 720p holds 12 x $5.778 = $69.34. If the same 24 slots are 10-second 480p drafts, they hold 24 x $2.68677 = $64.48, less than the half-full 720p queue.
The practical point is that a mixed queue of drafts and finals is cheap to hold, and a queue of finals is not. If you expect to run close to your balance, submit finals in smaller groups so that a failed submit does not strand a half-finished set.
For the plan limits themselves, read your workspace's generation_limits, because an admin override can differ from the defaults in the table.
Caveats
- Plan capacities are the documented defaults; admin overrides change them. Prefer
generation_limitsin a submit response. - Mixed queues reserve each job's own estimate. A queue of 480p drafts holds a fraction of the 1080p figure.
- I assume every accepted job reserves in full at submit, as the admission doc states for paid generation.
Using the ceiling in practice
Treat the figure as a ceiling for risk planning, not a forecast. Three uses are worth the arithmetic.
The figure scales with job length. A 5-second 1080p Seedance 2.5 job holds 5 x $1.42155 = $7.10775, so a full Scale queue of 120 holds $852.93, half of the 10-second figure. A 15-second job holds $21.32325 each, which makes 120 jobs $2,558.79. Duration is the cheapest lever for reducing exposure, because it changes every job in the queue at once.
- Sizing a top-up: if your largest realistic queue is 20 ten-second 1080p jobs, keep 20 x $14.2155 = $284.31 available, or submit in smaller groups.
- Setting an alert: warn when the balance falls under the reservation of one wave, not under zero.
- Choosing a plan: concurrency and queue size come from the plan, so a higher plan raises both throughput and the maximum amount that can be reserved at once.
Sources
Related posts
More in Pricing
- Gemini 3.8 Flash TTS free tier vs paid: test five lines, then budget
Google's pricing page lists Gemini 3.8 Flash TTS as free of charge on the free tier. Sume bills each TTS job, with a 1-cent catalog minimum.
- Gemini Omni Flash 1.1 price ladder: 4K is 10x the 360p rate
Omni Flash is $0.03 to $0.30 per second on fal; Sume bills list x 1.25. Price of the 8-second default, a 10-second max clip and a minute at each resolution.
- GPT Image 2.5 at 1024: xhigh is about 3,122 tokens and max about 7,024
Sume's docs put xhigh at $0.09366 and max at $0.21072 for a 1024 image at $30 per million tokens. That implies 3,122 and 7,024 tokens, before Sume's margin.
- GPT Image 2.5 quality auto holds the max price: a 12-image batch
Quality auto on gpt-image-2.5 (Flare, Sunburst) reserves max: $0.21072 of output at 1024x1024 before Sume pricing, so twelve images hold $3.16 or more.
Written by Sume