Veo 3.1 4K needs an 8-second clip; Sume's Veo rows stop at 720p
Google offers Veo 3.1 1080p and 4K only at 8 seconds. Sume's Veo rows run 720p at 4, 6 or 8 seconds. See the gap and which Sume row reaches 4K.

Google's Veo 3.1 gives 1080p and 4K only at 8 seconds, and Veo 3.1 Lite has no 4K at all (read 2026-10-10). Sume's Veo rows are fixed at 720p, so no Veo resolution above 720p is available through Sume.
If you want 4K from Sume, the row that lists it is gemini-omni-flash-1.1, which accepts 360p, 720p, 1080p and 4K.
Google's resolution and duration rules
The Gemini API Veo page (read 2026-10-10) lists durations of 4, 6 and 8 seconds, 720p by default, and 1080p and 4K only at 8 seconds. Aspect ratios are 16:9 and 9:16. Latency is 11 seconds to 6 minutes. The output carries a SynthID watermark.
So a 4-second Veo clip at Google is always 720p.
| Duration | 720p | 1080p | 4K |
|---|---|---|---|
| 4 s | Yes | No | No |
| 6 s | Yes | No | No |
| 8 s | Yes | Yes | Yes (not on Lite) |
What Sume's Veo rows do
Sume's Veo request builder accepts only 720p. Durations are 4, 6 and 8 seconds, with 16:9 or 9:16. The audio flag defaults to on. Billing is Google list times 1.25, so an 8-second 720p clip with audio bills $4.00 on veo-3.1, $1.00 on veo-3.1-fast and $0.50 on veo-3.1-lite.
These rows appear only when the environment switch is on and GET /v1/videos/models returns the id. Check the catalog before you commit.
Shorter clips cost proportionally less because the rate is per second. At the same 720p with audio, a 4-second clip bills $2.00 on veo-3.1, $0.50 on veo-3.1-fast and $0.25 on veo-3.1-lite, and a 6-second clip bills $3.00, $0.75 and $0.375. Turning audio off lowers the rate to $0.25, $0.10 and $0.0375 per second. Because the output is 720p at every length, the price is the only thing that changes between 4, 6 and 8 seconds, not the pixel count.
Rows on Sume that reach higher resolutions
Per the Video Generation docs, Gemini Omni Flash 1.1 goes to 4K for 3 to 10 seconds. At 4K, Sume bills $0.375 per second (list $0.30 times 1.25), so an 8-second 4K clip bills $3.00. Seedance 2.5, Wan 3.0 and Seedance 2 reach 1080p, and MiniMax H3 bills 2K and 4K upscales if the request includes them.
The table fixes the length at 8 seconds and the resolution at each row's highest. Prices are Sume billed rates, list times 1.25. Seedance 2.5 is the expensive end because it bills by video tokens, so its 1080p second costs far more than Wan's. Veo is shown at 720p because that is all Sume's rows run, and it appears only if your catalog lists it.
| Sume id | Top resolution | Per second | 8-second cost |
|---|---|---|---|
| gemini-omni-flash-1.1 | 4K | $0.375 | $3.00 |
| wan-3.0 | 1080p | $0.25 | $2.00 |
| seedance-2.5 | 1080p | $1.42155 | $11.3724 |
| veo-3.1 with audio (if listed) | 720p | $0.50 | $4.00 |
Reading the catalog and choosing
The supported_resolutions field in GET /v1/videos/models is the only dependable source. Send a resolution that is in that list for the row, or the request fails validation. Do not assume that a resolution Google supports at 8 seconds exists on Sume, because the Veo builder in the API code accepts 720p and nothing else.
For 1080p at Veo's 8-second length, Wan 3.0 is the inexpensive stand-in on Sume. For a 4K master, Omni is the only listed row. For the Veo look itself, and for any Veo output above 720p, use Google directly.
One more point on length: Google ties its higher resolutions to the 8-second setting, so a short 4-second Veo test tells you nothing about 1080p or 4K quality. On Sume, Omni accepts 4K at any length from 3 to 10 seconds, so you can test a 3-second 4K clip for $1.125 before you commit to a longer one.
Sources
Related posts
More in Models
- Veo 3.1 Lite (Preview) on Sume: what the preview label means
Sume's catalog names the row 'Veo 3.1 Lite (Preview)'. Its limits, the audio rates, and how it differs from the Gemini API's Lite lines for 720p and 1080p.
- Which Sume video row takes 1:1, 3:4 or 21:9? Aspect ratio matrix
Seedance and MiniMax rows take 21:9; Wan 3.0 skips it; Kling 3 stops at 16:9, 9:16 and 1:1; Omni and Veo are 16:9 or 9:16 only. The full matrix by row.
- An OpenRouter-compatible video API: sume/auto or a pinned model
Sume's POST /v1/videos follows OpenRouter's video generation API field for field. Let sume/auto pick the model, or pin a catalog id like seedance-2.5.
- Image generation API with reference images: POST /v1/images
Send a prompt plus public HTTPS reference images to Sume's POST /v1/images. Pin a catalog model or send sume/auto; the catalog lists each model's limits.
Written by Sume