12 s clip: sume/auto caps at 10 s; pin Wan $1.50 or Seedance 2.5 $6.94

You cannot get a 12-second clip from sume/auto: its create controls take 3 to 10 seconds, at 16:9 or 9:16, with defaults of 720p and 8 s. For 12 s, pin a catalog model. At 720p the pinned prices on Sume run from $0.90 on MiniMax H3 (native 768p) to $6.94 on Seedance 2.5.

Pinned prices for 12 seconds

Each row is a model that accepts 12 s. Gemini Omni Flash 1.1 is excluded because it stops at 10 s. Sume bills list x 1.25, rounded up to the cent.

What you give up by pinning

With sume/auto, the request body is the same and Sume selects the family, but the response always reports sume/auto and Sume never discloses which family ran. That is convenient until you need a length Auto does not offer. When you pin, you choose the family, the resolution and the aspect ratio yourself, so the price is knowable before you submit: for Wan 3.0 it is $0.10 x 12 x 1.25 = $1.50.

Two ways to reach 12 seconds

• One pinned job: choose from the table. Wan 3.0 accepts 2 to 30 s and Kling 3 accepts 4 to 15 s.
• Two Auto clips of 6 s each, joined on the timeline. The Sume docs do not publish an Auto price, so you cannot quote this route in advance, and the two clips may not match in look because Auto can choose a different family each time.

Recommendation

For one continuous 12 s shot, pin Wan 3.0 at 720p for $1.50, or H3 at 768p for $0.90 if the narrower pixel size is acceptable. If you want the Seedance 2.5 request shape with reference audio and 21:9 through 9:16 ratios, budget $6.94. Send a pinned model through Video generation (POST /v1/videos) or the Video Router, and keep Idempotency-Key on retries so a replay returns the original job rather than a second charge.

Check it before you run it

Every figure above is a catalog list price times 1.25, rounded up to the cent, as of 2026-10-08. Catalogs change, so before a large batch, read the current model entry in the docs and recompute the one line that matters for your case. Write the arithmetic next to the job in your own notes: list rate, seconds or characters, multiplier, rounding. If the result differs from the wallet charge by more than a cent, the catalog entry has changed, and the docs page is the place to find out why.

Run one small job first. Submit a single request with the pinned model id and the settings in the tables, poll the returned polling_url until it finishes, and compare the charge with your estimate. Then scale up. Using a pinned id for the test matters, because sume/auto never names the family, so you cannot tie its charge to the row you priced.