What does Sume's 25% cover? Container boot, idle tail and storage
The x1.25 on a Modal-priced job pays for container boot, the idle scaledown tail, min_containers floors and Volume storage. None of those is billed per job.

Sume does not add separate line items for container boot, the idle scaledown tail, min_containers floors or Volume storage. Those costs are absorbed by the **25% on top of list**. The per-job charge is container seconds x Modal list x 1.25, plus the 5.5% platform fee, clipped to the hold.
What is and is not billed per job
Modal bills the account for time containers run, including the time a container sits warm after work ends. A direct Modal user sees that in the bill as one pool of seconds. Sume customers do not have a Modal account, so the pool is Sume's.
| Cost | Billed to you per job? | Where it is covered |
|---|---|---|
| Container boot | No | The 25% margin |
| Idle scaledown tail | No | The 25% margin |
| min_containers floors | No | The 25% margin |
| Volume storage | No | The 25% margin |
| Container seconds used by the job | Yes | Modal list x 1.25, then +5.5% |
| Agent sandbox (E2B) | No second meter | The platform fee |
What the platform fee covers
E2B, the sandbox the agent uses, and the model sandbox have no second meter either. There is one fee, the 5.5% platform fee, and it covers them.
The arithmetic a Modal user would check
Modal's pricing page lists $0.0000131 per physical core-second and $0.00000222 per GiB-second. A Sume job's model price is those rates times the seconds it used, times 1.25. If you rent the same shape directly, you pay the list rate for the job, plus whatever idle tail and storage your account accumulates.
The honest comparison is therefore: direct is cheaper per busy second, and Sume is a flat 1.31875 multiplier that includes the idle time you would otherwise carry.
When this matters
For bursty media work, such as a probe here and a frame sample there, the idle tail can be larger than the job itself. For a steady workload that keeps a container busy, direct Modal is closer to its list price. If that is you, compare against the 31.875% total, not against the 25% alone.
Why Sume does not itemize it
Itemizing boot time and idle tails would make the bill depend on how busy the shared pool happened to be when your job ran. A flat multiplier on the published rate gives every customer the same price for the same job, whatever else is running.
The tradeoff is explicit: you pay for a job's busy seconds at 1.25x list, and you do not see or pay for the warm containers that make the next job start quickly.
Using this to estimate
If you want the cost of a Modal-priced job, use the Modal rates x the seconds x 1.31875. You do not need to guess an idle surcharge. The min_containers floors, which keep some containers warm so jobs start fast, are not a separate meter either.
Sume's published Modal card id is modal_list_2026-09-22_x1.25, so you can tell which rate card a usage row was priced from.
Sources
Related posts
More in Pricing
- Why a 3-second Gemini Omni clip bills 12 cents, not 11, on Sume
Sume rounds each video job up to the cent after the 1.25 multiple: a 3-second 360p Omni Flash clip is $0.12, a 3-second 4K clip $1.13. Row-by-row math.
- How Sume pricing works: plans, one wallet, published model rates
Sume plans set access and concurrency. Usage draws from one prepaid wallet at each model's published USD rate, for generation, the Agent, Formats, and the API.
- AI avatar video API pricing: cost per second and per minute
Sume bills AI avatar video per second by quality tier, with separate rates when you send a product image. Per-minute costs for standard, plus, and max.
- AI video generation cost per video: what one Sume API run cost
To see what one Sume run or video job cost, call GET /v1/usage with run_id or job_id and read debited_usd: the wallet deduction, agent turns included.
Written by Sume