Nano Banana 2 Lite first frame, then Omni Flash: the cost ladder

A cheap route to an Omni clip is a first frame from Nano Banana 2 Lite, listed by Google at $0.034 per 1K image, followed by an image-to-video call to Gemini Omni Flash at about $0.10 per second at 720p. Pick the still first, pay for video once. On Sume, send the still as image_url on gemini-omni-flash-1.1; video is billed at list x 1.25.

Prices come from Google's Nano Banana 2 Lite and Omni launch post and the Gemini API pricing page (read 2026-10-03). I do not state a Sume image price here, since this post only covers the video leg on Sume.

The ladder

Generate a few stills, choose one, then animate. Each extra still costs cents, while each extra video attempt costs dollars. The arithmetic for a 5 second 720p clip with four stills tried is below.

Keep the still and the clip consistent

Write the video prompt to continue what the still shows: same subject, same light, then the motion. Google lists limits on character consistency across scene changes and panning at launch, so keep the move small in the first test. Use end_image_url if you also have a last frame.

Send the still

The still must be a public HTTPS URL.

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: still-to-omni-001" \
  -d '{
    "model": "gemini-omni-flash-1.1",
    "prompt": "The woman turns toward the window and smiles. Same light, slow push in.",
    "image_url": "https://example.com/first-frame.png",
    "resolution": "720p",
    "duration": 5,
    "aspect_ratio": "9:16",
    "mode": "async"
  }'

On Sume

Sume's Video Router docs show the image-to-video fields. On Sume the video leg is 5 x 0.125 = $0.625 at 720p.