One 30-second Wan 3.0 job or four 8-second jobs: which costs less?

One 30-second Wan 3.0 clip at 720p costs $3.75 on Sume, and four 8-second clips cost $4.00 for 32 seconds. The rate per second is the same, $0.125, so the extra $0.25 is only the two spare seconds you paid for. The real difference between the two plans is continuity and what a bad take costs, not the price per second.

The arithmetic

Sume bills the provider list rate multiplied by 1.25 on every video model (Video Router). The figures here come from the price estimators in Sume's repository, run on 2026-10-07, and are rounded to the cent. The live numbers for your workspace are in the catalog response and on the API pricing page.

Wan 3.0 is billed as seconds times a resolution rate, with no per-job fee on the generation itself, and accepts 2 to 30 seconds (Video generation).

Where the plans really differ

A single job gives you one continuous shot from one prompt. Four jobs give you four starts, each with its own first frame, so you must match lighting and character across the joins. In exchange, a failed 8-second take costs $1.00 at 720p to redo, while a failed 30-second take costs $3.75.

If your acceptance rate for 8-second clips is 75 percent, the expected cost of a kept 8-second clip is $1.33 at 720p. If it is 50 percent for a 30-second clip, the expected cost of a kept 30-second clip is $7.50.

• Want one unbroken shot: choose the single 30-second job.
• Want editorial control per beat: four short jobs and a Timeline join.
• Want the cheapest total for exactly 30 seconds: three 10-second jobs match the 30-second price, with no spare seconds.

Reserve and concurrency

Sume reserves the estimated amount when it accepts a job, captures it on success, and refunds the reservation if the job fails or is cancelled before capture (Usage, Generation admission). So the balance you need on hand at submit time is the price of everything in flight, not only the final total.

The single job holds $3.75 for as long as it runs; four jobs hold $4.00 in total but can run in parallel, so they usually finish sooner on a plan with spare slots. A Pro workspace runs 4 jobs at once, Startup 8 and Scale 20 (the plans page lists the ladder), so a big batch drains in waves rather than all at once.

Set a cap before you start: from MCP, a paid tool takes dry_run=true and max_spend_usd (MCP tools and gates). Send a distinct Idempotency-Key per job so a retry replays the original job instead of charging again.

Check the figures against your own account

Every number on this page comes from the model rates and the 1.25 multiplier above, so you can reproduce it with a calculator. The live record is your own ledger: after a first job, read the balance and the usage rows for that job and compare the captured amount with the table. If they differ, trust the ledger and the model list your account returns, because the rates in the API response are the ones that bill.

The balance call below needs only an API key from your workspace and prints the USD balance. It does not start a job and costs nothing.

curl https://api.sume.com/v1/balance \
  -H "Authorization: Bearer $SUME_API_KEY"

What these figures leave out

The prices here cover video generation only. Voice, captions, a Timeline join and the plan fee are separate lines, and the API pricing page lists them. Provider list rates can change, so read the live catalog before you quote a client.