Google · video

Gemini Omni Flash

Google's omni video model: text, a start frame or up to 10 references, 3 to 10 seconds at 720p.

Text to videoImage to videoReferencesNative audio
Length
3 to 10 s
Aspect ratio
16:9, 9:16
References
Up to 10 image files

Opens the Studio with Gemini Omni Flash and your prompt. You see the price before anything runs.

About Gemini Omni Flash

Gemini Omni Flash is Google's omni video model, priced at $0.33 per second. It runs as text to video, image to video and references to video, with audio generated with the picture, from 3 seconds to 10 seconds, at 16:9 and 9:16.

The provider bills it by tokens, which works out to a rate per second plus a little for each reference image. A typical generation takes about 2 min.

Good for

  • Clips with synchronized audio that rely on real-world knowledge and physical interaction, which the provider says the model is grounded in
  • Combining up to 10 reference images in a shot

Getting good results

  1. The provider's prompt guide lets you bind each reference image to a role in the prompt with a numbered image tag.
  2. Each reference image adds a small amount to the price, on top of the rate per second; the clip itself costs $2.60 at the default length.

Gemini Omni Flash questions

What is the difference with Gemini Omni Flash 1.1?

Gemini Omni Flash 1.1 adds a start-and-end-frame mode, reference videos and a resolution choice from 360p to 4K, and starts at $0.075 per second. Gemini Omni Flash has no resolution setting and costs $0.33 per second.

Other video models

AI video generator