4K vertical 9:16 vs square 1:1 on Sume: which video models take which
Omni takes 4K in 16:9 or 9:16 but not square. Kling 3 and Seedance 2.5 take 1:1 but top out at 1080p on Sume. Pick the shape first, then the model.

If you need 4K vertical, use gemini-omni-flash-1.1, which takes 16:9 and 9:16 up to 4K. If you need a square 1:1 clip, kling-3 and seedance-2.5 take it, but their top resolution on Sume is 1080p, not 4K.
There is no row that does both on Sume today, so choose between a bigger frame and a different shape. The numbers below come from the catalog and docs in the repo.
The shapes each model accepts
Omni's envelope is 360p, 720p, 1080p and 4K, with 16:9 or 9:16 only. Auto uses the same envelope, because its default target is Omni, so an auto request for a 1:1 clip fails closed with a 400 unsupported_capability. Auto never silently routes to Seedance for it.
For a square clip, name kling-3 or seedance-2.5 in the request.
| Model id | 16:9 and 9:16 | 1:1 | Top resolution on Sume |
|---|---|---|---|
| gemini-omni-flash-1.1 | yes | no | 4K |
| kling-3 | yes | yes | 1080p |
| seedance-2.5 | yes | yes | 1080p |
| sume/auto | as Omni | no, 400 | 4K |
What a 4K vertical clip costs
Omni 4K is $0.30 a second at provider list, so $0.375 on Sume with the 1.25 multiple. A 10-second 4K clip is 10 x $0.375 = $3.75. The same clip at 1080p is 10 x $0.1875 = $1.875, which rounds up to $1.88.
A 4K clip costs exactly twice a 1080p clip, since $0.375 / $0.1875 = 2.
A vertical 4K request
Send aspect_ratio: 9:16 and resolution: 4K. If you are unsure about a prompt, draft at 720p first, since a draft costs a third of the final.
curl -X POST https://api.sume.com/v1/videos \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: vertical-4k-001" \
-d '{"model":"gemini-omni-flash-1.1","prompt":"A barista pours latte art, slow push-in, warm morning light","duration":10,"resolution":"4K","aspect_ratio":"9:16"}'If you need square and sharp
Generate square at 1080p with kling-3 or seedance-2.5, then upscale it with a tool of your choice if the placement needs more pixels. Check the Sume docs for the current video tools before you plan around one.
For a reel or a short, 9:16 at 1080p is the usual target, so most vertical work does not need 4K at all.
- Decide the shape first, because it limits the model.
- Draft at 720p, then render the final.
- Do not rely on
sume/autofor 1:1, since it returns a 400. - Read the catalog if you are unsure, as shown in the catalog filter script.
Related posts
More in Models
- '720p is not supported by model minimax-h3; use 768p' and its twins
MiniMax H3 and H3 Max are 768p models on Sume: 720p gets a 400 that says use 768p. 768p is refused elsewhere, and 360p is Gemini Omni only.
- 9:16 cover still: 1080x1920 fails GPT Image 2.5, 1088x1920 passes
GPT Image 2.5 needs edges in multiples of 16, 655,360 to 8,294,400 pixels, and a ratio of 3:1 or less. 1088x1920 and 1152x2048 pass; 576x1024 is too small.
- AI blueprint-style poster: Ideogram 4.5 at $0.0375, $0.075 or $0.275
Ideogram 4.5 bills $0.0375, $0.075 or $0.275 per image on Sume by quality (list $0.03, $0.06, $0.22), the same at 1K and 2K; draft low, finish high.
- AI book cover mockup on Nano Banana Pro: 4:5 and four resolutions
Nano Banana Pro lists at $0.15 at 1K and default quality and bills $0.1875 on Sume; it takes 0.5K, 1K, 2K or 4K and 4:5, so read usage.cost for larger tiers.
Written by Sume