3-second logo sting with Omni: end frame as the lockup, cost on Sume

Make a 3-second logo reveal: first frame a texture, last frame your logo lockup. Omni 3 s is $0.12 at 360p and $0.57 at 1080p on Sume. Request and checks.

5 min readSume
All posts

To make a three-second logo sting, send a plain texture or scene as the first frame and the finished logo lockup as the last frame to gemini-omni-flash-1.1. A 3-second clip is $0.12 at 360p, $0.38 at 720p and $0.57 at 1080p on Sume, and three seconds is the shortest length the model accepts.

First and last frame control is an Omni 1.1 Flash feature that generates continuous video between two keyframes (Google blog, read 2026-10-05). The last frame is where a logo benefit shows up: you control the exact lockup the viewer sees at the end.

Why the end frame matters

Text-to-video models draw logos approximately. A last frame fixes the final pixels of the clip to your art, so the sting always ends on the right wordmark. Make that frame a clean lockup at the target aspect ratio, 16:9 or 9:16, on a plain background with generous margin so the model has room to arrive at it.

  • First frame: a texture, a color field, a blurred shot of the product, or an empty version of the lockup background.
  • Last frame: the lockup exactly as brand rules specify.
  • Prompt: the move between them, for example ink spreading, light sweeping, or the camera pulling back.
  • Do not ask the prompt to write text. The last frame already has it.

Cost across the ladder

A sting is short and gets re-run many times for approval, so the 360p draft at $0.12 is the sensible first step. Once the move is approved, one final at your delivery resolution finishes the job.

Omni 3 s sting price by resolution on Sume (computed from Sume rates, read 2026-10-05)
Resolution3 s10 per weekPer month (40)
360p$0.12$1.20$4.80
720p$0.38$3.80$15.20
1080p$0.57$5.70$22.80
4K$1.13$11.30$45.20

Request

Use the Video Router with image_url as the first frame and end_image_url as the last.

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: sting-brand-v3-draft" \
  -d '{
    "model": "gemini-omni-flash-1.1",
    "prompt": "Light sweeps across brushed metal and resolves into the logo, camera steady",
    "image_url": "https://example.com/metal.png",
    "end_image_url": "https://example.com/lockup.png",
    "resolution": "360p",
    "duration": 3,
    "aspect_ratio": "16:9",
    "mode": "async"
  }'

What three seconds can carry

Three seconds is short: about one beat. A sting that tries two moves reads as noise. Choose one verb for the clip, such as reveal, assemble, sweep or settle, and let the start frame set it up. Because the clip is cheap, test five or six verbs at 360p ($0.12 each, $0.72 for six) and keep the one that looks right. The prompt can be as short as one sentence.

The shortest valid Omni length is 3 seconds, so the sting is also the cheapest finished clip you can make with this model.

Variants of one sting

Brands usually need the sting in both 16:9 and 9:16. Omni takes both, so make two lockup frames and two requests with the matching aspect_ratio. At 1080p that is $1.14 for the pair. Seasonal variants only change the first frame, which keeps the lockup consistent.

Using a sting inside a longer video

A sting usually sits at the head or tail of another clip. Sume's Timeline API assembles ordered video slots over one audio spine and returns one MP4, with inputs that are already media in your workspace. A 3-second sting plus a 10-second Omni clip is $0.57 plus $1.88 at 1080p, $2.45 of generation before the assembly.

Prompt examples

Three starting prompts for the move between the frames: "Ink bleeds across white paper and resolves into the lockup, camera locked off." "The camera pulls back from a macro of the metal surface until the logo is centered." "Particles gather from the edges toward the center and settle into the logo." Each describes one verb and one camera behavior, and none mentions the wordmark text.

Run each at 360p for $0.12, pick the best, and finish at the delivery resolution.

Checks before you publish

Open the final at full size and check the last frame against the lockup file. Native audio is always on for this model, so the clip carries generated sound; if your brand has a sound logo, replace the audio on the sting rather than relying on the generated track. A Timeline can place the sting at the head or tail of a longer film.

Sources

Related posts

More in Use cases

All Use cases posts

Written by Sume