Gemini Omni Flash edits 10-second inputs; Sume uses video_url to edit
Google's Gemini Omni Flash edits video inputs up to 10 seconds. Sume routes gemini-omni-flash-1.1 edits through a video_url, with no duration or aspect_ratio.

Google's Gemini Omni Flash page says video input for edit and extend can be up to 10 seconds, and outputs run 3-10 seconds. Sume's catalog id is gemini-omni-flash-1.1, which is the same family with the number in a different place from Google's stable id gemini-omni-1.1-flash. Copying the id between the two will fail. In Sume, an edit is a request with a video_url source and a prompt that gives the instructions.
The two pages side by side
| Item | Google page | Sume docs |
|---|---|---|
| Model id | Stable gemini-omni-1.1-flash; preview gemini-omni-flash-preview | gemini-omni-flash-1.1 |
| Inputs | Text, image, video | Image and video references; no audio reference |
| Edit source | Video up to 10 seconds (edit/extend) | video_url; the docs give no input length cap |
| Output length | 3-10 seconds | 3-10 seconds (edit: the source sets the length) |
| Resolution | 360p, 720p, 1080p, 4K at 24 FPS | 360p, 720p, 1080p, 4K; edit resolution optional, default 720p |
| Context | 1,048,576 tokens | Not applicable |
An edit on Sume
Per the docs, video_url is the edit source and not a reference. You cannot send it together with image_url, end_image_url or reference_*_urls, and an edit takes no aspect_ratio (400 if sent); a duration is only a reserve hint. Native audio is always on, and the API rejects generate_audio: false. Because the Sume docs state no input length limit, trim your source to 10 seconds or less to stay inside the figure Google publishes. Video jobs are submitted asynchronously, so the example sets mode to async.
curl -X POST https://api.sume.com/v1/video-router/generate \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: video-router-edit-001" \
-d '{
"model": "gemini-omni-flash-1.1",
"prompt": "Replace the bottle with an apple. Keep everything else the same.",
"video_url": "https://example.com/clip.mp4",
"resolution": "720p",
"mode": "async"
}'What I could not confirm, and one id note
- Whether Sume supports Google's "extend" mode. The Sume docs describe edit (
video_to_video) only. - Whether Sume enforces the 10-second input figure. Check the live OpenAPI under Video Router.
- Google deprecated the preview id on 2026-09-30 after general availability on 2026-08-27; use stable ids in anything new. Sume's docs use
gemini-omni-flash-1.1.
Trimming and cost planning
Trim before you upload: a 24 second source needs a cut to 10 seconds or less to match Google's published limit, and Sume documents a video trim tool for that. Sume bills the provider list price times 1.25 for each output second, depending on the resolution.
An edit sends no duration to the provider, so the output length follows the source, within the 3 to 10 second model range stated on both pages. If you need more than 10 seconds, plan separate edits and join them with the timeline tools rather than expecting one longer call.
Sources
Related posts
More in Models
- Gemini Omni Flash GA on Aug 27: extension, interpolation, 4K on Sume
Omni Flash 1.1 went GA on August 27, 2026 with extension, interpolation and 360p to 4K. Sume exposes interpolation and resolution, but no extend parameter.
- Gemini Omni reference errors: 10 images, 3 clips of 3 s, 16:9 or 9:16
Gemini Omni Flash 1.1 on Sume refuses an 11th image, a 4th or longer-than-3-second clip, audio references, bitrate_mode and ratios other than 16:9 or 9:16.
- Gemini Omni video in the Windows app, and by API on Sume
Google lists Gemini Omni video in the Gemini app for Windows. For your own app, Sume exposes gemini-omni-flash-1.1 at 360p to 4K, 3 to 10 seconds.
- generate_audio false on Gemini Omni: accepted, clip keeps its audio
Video Router accepts generate_audio false on gemini-omni-flash-1.1 and ignores it; MiniMax H3 answers 400. Check the model before you rely on a mute flag.
Written by Sume