Gemini Omni 1.1 Flash 360p drafts: what a cheap preview costs

Yes: Gemini Omni 1.1 Flash can render a low-resolution 360p draft before you pay for the final. On Sume, gemini-omni-flash-1.1 accepts resolution: "360p", and a 360p second is billed at 30% of a 720p second.

As of 2026-09-29, Google's launch post for Gemini Omni 1.1 Flash says 360p previews cost a third of its standard 720p resolution. Sume's numbers below come from its Video generation docs and pricing code.

What does a 360p draft cost on Sume?

Sume bills video at the provider's list price × 1.25, reserved when you submit, plus a 5.5% agent fee by default. The table shows the rate per output second and a 5- and 10-second clip at 16:9.

How do I request a 360p draft?

Send the same request you would for the final, with resolution set to 360p. Only gemini-omni-flash-1.1 accepts 360p and native 4K, per the API reference; the other catalog models do not list 360p.

curl -X POST "https://api.sume.com/v1/videos" \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: omni-draft-001" \
  -d '{
    "model": "gemini-omni-flash-1.1",
    "prompt": "A slow push-in on a ceramic mug on a wooden desk, morning light",
    "resolution": "360p",
    "aspect_ratio": "16:9",
    "duration": 5
  }'

Will the final match the draft?

Not as a guarantee. A final is a new job: you resubmit the prompt at 720p, 1080p or 4K. Sume's docs say no v1 video model accepts seed, so there is no field that pins the result, and the final may differ from the draft. Use drafts to test the prompt and framing, then judge the final on its own.

How do I get the draft back?

A draft is a normal video job, so the flow is the one every model shares: submit to POST /v1/videos, receive a job id and polling URL, poll GET /v1/videos/{jobId} until the status is completed, then download from the content URL. A job that ends failed carries an error field, and a cancelled job was canceled before it finished.

Send an Idempotency-Key on each submit, as in the sample. Docs say a replay with the same key returns the original job, so a retry after a network error does not start a second draft. Use a new key for each new draft.

What stays the same between draft and final?

• Length: 3–10 seconds, default 8 in the API reference.
• Shape: 16:9 or 9:16.
• Audio: native synced audio is always on, and generate_audio: false is rejected.
• Inputs: image and video references, and the video edit path, follow the same model rules; see Gemini Omni Flash 1.1 API.

Should I draft at 360p or just render the final?

Compare the two lines of the table for your clip length. If you expect to rerun a prompt several times, cheaper drafts add up; if the first try is usually right, they only add a step. The arithmetic is yours to run with the rates above; How much does an AI video cost? shows the same method across models.