Omni Flash edit on Sume: what editing a 6 s and a 30 s clip costs
Editing a clip with Omni Flash on Sume bills per output second at the resolution you pick: $0.75 for 6 s at 720p, up to $3.75 for a 30 s hint. How it is priced.

Editing a clip with Gemini Omni Flash on Sume is billed per output second at the same resolution rates as generation, so a 6-second edit at 720p is $0.75 and a 30-second one is $3.75. The docs say edit output length follows the source clip, and the duration you send is only a hint used for the reserve.
That makes the edit price easy to predict if you know the clip length and the resolution.
How edit mode works
Per Sume's video router docs, you send a prompt and a video_url, resolution is optional and defaults to 720p, and you do not send aspect_ratio. The model's model card says edit is the only family in the catalog with video_url edit.
The duration field, if sent, is not forwarded. It tells Sume how much to reserve, up to a ceiling of 30 seconds.
| Source length | 360p | 720p | 1080p | 4K |
|---|---|---|---|---|
| 6 seconds | $0.225 | $0.75 | $1.125 | $2.25 |
| 10 seconds | $0.375 | $1.25 | $1.875 | $3.75 |
| 30 seconds | $1.125 | $3.75 | $5.625 | $11.25 |
What to watch
The table multiplies the per-second rate by the source length. We did not verify in the docs how an over- or under-sized duration hint is settled, so send a hint close to the real length and compare the final amount on the job.
Resolution is the main cost lever. Editing at 720p costs a third of 4K for the same clip.
When editing beats regenerating
If the shot is right and one element is wrong, an edit keeps the rest and costs the same per second as a fresh generation. That is only cheaper if it avoids further rerolls. A regeneration with a changed prompt is a different sample, not a patch of the old one.
Edits suit changes like replacing an object or altering weather. They suit worse the cases where the whole motion needs to change.
A cost check you can run
Before an edit job, write down the source length and the resolution you will request, multiply, and compare the result with the amount on the finished job. If a source clip is 12 seconds and you edit at 1080p, expect about $2.25, from 12 times $0.1875.
Do this on the first three edits of a new workflow. If the numbers match, you can budget the rest of the batch from the table. If they differ, read the job's usage fields before running more.
Google's side
Google's pricing page lists Omni Flash video by output tokens. We did not read an edit-specific price there, so no direct comparison is made. Sume's rates are the provider list times 1.25.
Sources
Related posts
More in Models
- Retired AI model id to Sume id: GPT Image, Gemini and Sora lookup
One lookup table that maps retired gpt-image-1, gemini-2.5-flash-image and sora-2 ids to Sume ids, with dates, so a failing call tells you what to switch to.
- Seedance 2.5 API: Seed blog says soon, BytePlus says live
Two ByteDance pages disagree on Seedance 2.5 API access. Here is what each says, and the seedance-2.5 id that runs on Sume's video router today.
- Seedance 2.5 reference-to-video: what Sume accepts per pass
ByteDance lists up to 30 images, 10 videos and 10 audio clips per pass for Seedance 2.5. Here is how that maps to Sume's seedance-2.5 reference types.
- seedream-v4 or bytedance-seed/seedream-4: which Sume id to store
Sume accepts legacy bare ids like seedream-v4 and seedream-4-5-edit, but the catalog lists org/slug ids. Store the canonical id; Python normalizer inside.
Written by Sume