Pin Gemini Omni Flash 1.1 or send sume/auto: six checks

Send sume/auto when you only need a clip of 3 to 10 seconds at 16:9 or 9:16 and do not care which model makes it; pin gemini-omni-flash-1.1 when the request depends on something specific to Omni: edit mode through video_url, the 4K tier, or the <IMAGE_REF_0> tags. The docs name Omni Flash as the Auto default for video, and Auto answers with sume/auto and never tells you which family ran.

Six checks

Go through the rows from the top. The first one that applies decides.

What Auto gives you

Auto resolves from the normalized request and the catalog version, so an idempotent replay gets the same price and the same route. The poll response says sume/auto. Sume does not disclose the family and the docs tell you not to guess it from the output.

That is useful for a pipeline that wants no model names. It is a poor fit for reporting that must say which model produced which clip.

{
  "model": "sume/auto",
  "prompt": "A vertical UGC-style product clip on a desk, natural light",
  "aspect_ratio": "9:16",
  "duration": 5
}

Cost does not decide it

On Omni, 5 s at 720p is 63 cents (62.5 rounded up) and 8 s is $1.00. The same prices apply whether you pin or use Auto when Auto uses Omni, so the choice is about control and reporting, not about saving money. If you need the cheapest possible draft, set resolution to 360p, which is 3.75 cents a second.

The same length at every tier

For reference, a 8-second Omni Flash clip at each resolution. Every price is the seconds times the billable rate, rounded up to a whole cent, as of 2026-10-08.

Limits to remember

These apply to every request on this page, from the Video Router and Video generation docs:

• Length is 3 to 10 whole seconds in generation modes; an edit takes no duration.
• Aspect ratio is 16:9 or 9:16; an edit takes no aspect ratio.
• Native synced audio is always on, and generate_audio: false is rejected.
• There is no bitrate_mode, no reference audio and no seed.
• Billing is the provider list times 1.25 per output second, reserved at submit and shown in usage.cost.