What is Gemini Omni 1.2 Flash?
Gemini Omni 1.2 Flash is Google's AI video generator for short clips with native sound. You describe a scene in plain sentences, and it renders one continuous take of 3 to 15 seconds, from a 360p draft up to 4K. On OpenClips it runs in the app, through the REST API and through the MCP server, and every MP4 lands in your library.
Text to video, image to video and reference to video
Omni 1.2 Flash works from what you give it:
- Text to video: a few plain sentences about the place, the action, the camera and the sound.
- Image to video: one image as the opening frame, or two images as the first and last frame of a 3 to 10 second transition.
- Reference to video: up to 6 images to keep a person, an outfit or a product the same, plus up to 3 short reference clips to guide the look and motion.
What is new in Omni 1.2 Flash
Omni 1.2 Flash makes longer clips than Omni Flash 1.1: up to 15 seconds instead of 10. It also takes short reference clips beside images, and the length you choose now decides the clip, whatever the prompt says. Google prices 1.2 lower than 1.1.
Edit and extend your Omni videos
Two companion models work on a clip you already have:
- [Gemini Omni 1.2 Flash Edit](/models/gemini-omni-flash-1-2-edit/) changes a clip of up to 15 seconds from a plain-words prompt, such as a new season, new clothes or a new set, and keeps its length, movement and framing.
- [Gemini Omni 1.2 Flash Extend](/models/gemini-omni-flash-1-2-extend/) adds 3 to 10 seconds to the end of a clip and continues the scene, the motion and the sound.
Prompt guide for Gemini Omni 1.2 Flash
Write the scene as you would brief a camera crew. Name the place and the light, say who or what is in frame, give one camera move, write how the shot ends, and name the sounds. Omni reads a short, plain description better than a long timeline. Keep exact words, prices and logos for the editor, where they stay sharp.