Sume Auto Video or Pin a Model? When sume/auto Is Enough

model: "sume/auto" on POST /v1/videos lets Sume choose the family. It is the lowest-effort route and the right one for a lot of shots, but it has fixed edges, and some requests need a pinned model.

What Auto guarantees

Auto create controls default to 720p and 8 seconds and accept 3 to 10 seconds at 16:9 or 9:16. Resolution is a pure function of the normalized request and the catalog version, so an idempotent replay prices and routes the same way. The poll response reports sume/auto as the model, and Sume does not disclose which family served the request. The docs say not to build on any observable trait of the output to infer it.

Billing follows the same rule as every catalog model: provider list times 1.25.

When to pin

Pin a model in these cases.

• The shot is longer than 10 seconds: seedance-2.5 and wan-3.0 reach 30.
• You need 1:1, 21:9 or another ratio besides 16:9 and 9:16.
• You need audio or video references, a person swap or a motion transfer.
• You need a fixed price per second for a budget, such as wan-3.0 at $0.125 at 720p.
• You need the same family across a series, because Auto does not tell you which family ran.

A simple policy

Use Auto for exploration and for one-off clips where 8 seconds at 720p is the deliverable. Move to a pinned id as soon as a shot becomes part of a series, a template or a client deliverable. GET /v1/videos/models lists what you can pin and each model's limits.

Both routes create the same kind of job, visible at GET /v1/jobs/{id}/status, so switching later is a one-field change. See which model fits the inputs you have.