Gemini Omni 4K in the Sume API: send 4K or 4k? Both work
Sume accepts 4K or lowercase 4k for Gemini Omni Flash 1.1 and no other model. 4K costs $0.375 per second, so a 10 s clip is $3.75. Limits and a request.
For Gemini Omni Flash 1.1 on Sume, send resolution: "4K" or lowercase 4k. Both work, and Sume translates to the provider's lowercase enum. 4K costs $0.375 per second, so 10 s is $3.75. The lowercase alias is accepted only on this model. Wan 3.0 and Seedance 2.5 stop at 1080p.
Why two spellings
The catalog token for the top tier is 4K, shared with the H3 upscale SKU. The provider's input enum is lowercase 4k, so the payload builder translates (a comment in pricing-tables.ts says so). The Video Router docs add that only this model accepts lowercase 4k as an alias. Elsewhere, use the catalog spelling.
What 4K costs
Omni lists $0.30 per second at 4K. At the house x 1.25 that is $0.375. The default resolution is 720p, so you must ask for 4K.
| Duration | 720p | 1080p | 4K |
|---|---|---|---|
| 3 s | $0.38 | $0.57 | $1.13 |
| 5 s | $0.63 | $0.94 | $1.88 |
| 8 s | $1.00 | $1.50 | $3.00 |
| 10 s | $1.25 | $1.88 | $3.75 |
Limits at 4K
Omni runs 3 to 10 seconds in 16:9 or 9:16 only. Native synced audio is always on, and generate_audio: false returns 400. The Agents Videos page does not offer 4K under Auto; it offers 720p and 1080p there, so 4K is an API option.
Request
Draft at 720p first: a 10 s draft is $1.25 against $3.75 at 4K. See our 4K model list for every tier.
The job is async, so poll polling_url and download unsigned_urls[0] when status is completed.
curl -X POST https://api.sume.com/v1/videos \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"gemini-omni-flash-1.1","prompt":"Macro shot of a watch mechanism, shallow focus","duration":10,"resolution":"4K","aspect_ratio":"16:9"}'Draft low, finish high
Omni is a good candidate for a two-step flow because the tiers share one request shape. Run the prompt at 360p ($0.0375 per second, 3 s for $0.12) to check composition, 720p to check motion and sound, and 4K for the delivery file. There is no seed, so the 4K take will differ from the draft; use the draft to approve the prompt, not the frame.
If the 4K result is for a client file, keep the job id, and download from unsigned_urls[0] as soon as the job completes.
Related posts
More in Developers
- Gemini Omni reference clips: 3 videos, 3 seconds each, VIDEO_REF_0
On Sume, Gemini Omni Flash reference-to-video takes up to 10 images and 3 clips of at most 3 s each, addressed as IMAGE_REF_0 and VIDEO_REF_0 in the prompt.
- Gemini Omni Flash edit on Sume: the 8-second reserve hint
A Gemini Omni edit has no duration input, so Sume reserves against a hint: 8 seconds by default, up to 30 if you send one. At 720p, that is $1.00 held.
- Gemini Omni Interactions API versus Sume /v1/videos: field mapping
Google returns Omni video as base64 in an Interactions response. Sume returns 202 with a polling_url and a content URL. See the mapping and a bytes helper.
- Gemini Omni prompt template: five parts in a Python builder
Google's Omni guide names five prompt parts: framing, style, lighting, location, action. Build them in Python and send the prompt to Sume's /v1/videos.
Written by Sume