Omni Flash

Google’s Gemini Omni Flash: video with sound at 720p and 1080p, 6 or 8 seconds long, from a prompt, a start image, a start and end image, or up to 3 reference images.

Prices

A clip costs its length times the rate: a 6 second 720p clip is $0.282. It’s charged only when the video is delivered.

What it takes

• A prompt of up to 8,000 characters.
• A first_frame (and optionally a last_frame), or up to 3 reference_images, not both.
• Lengths of 6 or 8 seconds.
• 16:9 and 9:16. From a first frame, set aspect_ratio to the image’s shape.
• Sound is always on.
• 720p and 1080p.

Example

{
  "model": "gemini-omni-flash",
  "first_frame": "https://example.com/river-valley.jpg",
  "last_frame": "https://example.com/river-valley-night.jpg",
  "prompt": "A timelapse of the river valley from sunrise into a clear starry night; the drone holds its position as the farmhouse windows light up.",
  "seconds": 8,
  "resolution": "720p",
  "aspect_ratio": "16:9"
}

Prompting

• Describe the scene, the motion, the camera, the light and the sound: Omni makes the audio too.
• Time the action within the clip if it matters: “[0-3s] she walks in, [3-6s] she turns to the window”.
• Say what to leave out in the prompt itself (“no text on screen”): there’s no negative_prompt.

Good to know

• Clips are usually ready in one to two minutes.
• Frames are passed to the model as images with Google’s own instruction to start (or end) on them, as Google’s Omni API does.
• There’s no 360p, 4K, video editing or extension yet.
• Prompts or images that may break the content rules can be declined (content_policy); declined videos are free.
• Videos are kept for 30 days; download the ones you want to keep.

Full API details: Video generation. Try it in the dashboard’s Playground, which shows the code for each video.