Gemini Omni Flash 1.1 API: a vertical 4K clip with curl on Sume

Request a 9:16 4K clip from Gemini Omni Flash 1.1 on Sume: model id, resolution 4K, 3 to 10 seconds, audio always on, and the $3.00 price for 8 seconds.

4 min readSume
All posts

To get a 4K clip from Gemini Omni Flash 1.1 on Sume, POST model: "gemini-omni-flash-1.1" with resolution: "4K", aspect_ratio of 9:16 or 16:9, and a duration between 3 and 10 seconds. An 8-second 4K clip costs $3.00.

Sound is always generated; there is no audio switch, and generate_audio: false is a 400.

Request

Set your key in SUME_API_KEY. The job returns 202 immediately and you poll the polling_url it gives you.

curl -X POST "https://api.sume.com/v1/videos" \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: omni-4k-001" \
  -d '{"model":"gemini-omni-flash-1.1","prompt":"A glass perfume bottle on wet black stone, slow orbit, soft rim light","duration":8,"resolution":"4K","aspect_ratio":"9:16"}'

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

Price ladder

Sume prices Omni Flash per second by resolution at the provider list times 1.25, rounded up to whole cents. The default resolution when you omit the field is 720p.

Gemini Omni Flash 1.1 on Sume, eight seconds (read 2026-10-07)
ResolutionProvider list per secondSume per second8 seconds
360p$0.03$0.0375$0.30
720p (default)$0.10$0.125$1.00
1080p$0.15$0.1875$1.50
4K$0.30$0.375$3.00

Vendor context

Google's Gemini API video page presents Gemini Omni Flash as a fast multimodal model for video generation and conversational editing, and recommends it as the default unless you need capabilities such as scene extension or frame control, which it assigns to Veo 3.1 (read 2026-10-07). Sume lists Omni Flash 1.1; it does not list Veo 3.1.

The default model for sume/auto is also Omni Flash 1.1, with a 3 to 10 second range, 720p default and 8-second default length. Pin the id as above when you want 4K, because the other models have different ceilings.

Details people miss

The API accepts 4K and also lowercase 4k on this model only. Image-to-video uses frame_images; a single image placed in input_references is reference-to-video, not a first frame. Reference-to-video takes up to 10 images and 3 video clips of at most 3 seconds each.

There is no extend field and no seed. If you want to change a clip you already have, use the edit mode through the Video Router with video_url instead of regenerating.

  • Ratios: 16:9 and 9:16 only.
  • Duration: 3 to 10 seconds, whole numbers.
  • No bitrate_mode and no audio reference URLs.

Sources

Related posts

More in Developers

All Developers posts

Written by Sume