Sneaker drop teaser in 6 seconds: Gemini Omni timecodes and cost

A sneaker teaser fits in 6 seconds: one product photo as the first frame, a timecoded prompt with two beats (reveal, then hero angle), and a bass-hit sound line. On Sume, Gemini Omni at 720p costs $0.75 for the clip and $1.13 at 1080p. Google's Veo 3.1 allows only 4, 6 or 8 seconds (read 2026-10-07), so 6 seconds is also the length you would use there; Sume does not list Veo.

Keep the shoe the hero: no people, one surface, one light.

Two beats, one move each

Google's Omni page gives timecode syntax in the form '[0-3s] ..., [3-6s] ...' (read 2026-10-07). Use it with one camera move per range.

• [0-3s] 'The shoe sits in darkness; a rim light sweeps across the upper, camera low and close.'
• [3-6s] 'The camera rises to a three-quarter hero angle as the sole lights up. Hold.'
• Audio: 'a deep bass hit at the reveal, a soft whoosh on the camera rise. No dialogue.'
• Do not ask for brand text unless it is on your first-frame photo.

Request

The photo goes in as first_frame, which keeps the shoe's design from being reinvented.

{
  "model": "gemini-omni-flash-1.1",
  "prompt": "Single continuous product teaser. [0-3s] the shoe sits in darkness, a rim light sweeps across the upper, camera low and close. [3-6s] the camera rises to a three-quarter hero angle as the sole lights up, hold. Audio: deep bass hit at the reveal, soft whoosh on the rise. No dialogue.",
  "duration": 6,
  "resolution": "720p",
  "aspect_ratio": "9:16",
  "frame_images": [
    { "type": "image_url", "image_url": { "url": "https://example.com/sneaker.jpg" }, "frame_type": "first_frame" }
  ]
}

Cost against Google's list prices

Sume bills list times 1.25 per second, rounded up to the cent per job. The Veo rows are Google's list per second from its pricing page, times 6, and are shown for scale only.

What to verify

Compare the shoe in the last frame with your photo: laces, stripes and sole color. If the design changes, shorten to 4 seconds and remove one camera move. The bass hit is only a request to the model, so check that it lands near the 3-second mark. If you want a brand track, you can replace the audio in a Sume timeline, at the cost of an extra render.

Variants on a list

A drop usually has several colourways. Keep the prompt unchanged and swap only the first-frame photo for each colourway; the identical timecodes and sound line make the variants look like a family. Submit them with distinct Idempotency-Keys such as sneaker-red-v1 and sneaker-blue-v1, so a retry never double-charges. At 6 seconds and 720p, ten colourways cost $7.50 on Sume before any re-rolls, and a full set at 1080p costs $11.30 from the rounded per-job price of $1.13.