← All models
GoogleVideo

Gemini Omni Flash

High-performance multimodal video generation and editing

Try Gemini Omni Flash

About

Gemini Omni Flash is Google's multimodal video model that accepts text, images, and even existing video as input. It can generate video from scratch, animate reference images, or edit existing footage — all through natural language instructions. With support for up to 9 reference images and direct video input, it's the most flexible video model in the Lyvia lineup.

CinematicCreativeExperimentalCommercial

How to prompt

  • You can provide a video as input and describe the edit: "make it look like a watercolor painting", "change the sky to sunset"
  • Combine reference images with text for precise direction: use images for style, text for action
  • Keep prompts focused on one clear action per generation for best coherence
  • Use 16:9 for cinematic scenes, 9:16 for social media content

Best for

Video editing, style transfer, multimodal creative workflows, animating reference images, experimental video

Good to know

Limited to 16:9 and 9:16 aspect ratios. No duration control — output length is model-determined.

Ready to create with Gemini Omni Flash?

Open Lyvia