Produce High-Speed Conversational Video with Gemini Omni Flash
Create and edit 720p videos using high-speed multimodal models.
Switch to 'gemini-omni-flash-preview' via the Interactions API to generate and conversationally edit 3-10 second videos.
The Scenario
You need to create a short, professional video clip or animate a still product image for a social media post or presentation.
Before & after
Creating 10-second high-quality video animations manually using professional editing software typically takes 2-4 hours of design and rendering.
Use Gemini Omni Flash with the Interactions API to generate and iteratively refine video content through simple prompts in 1-2 minutes per iteration.
The Prompt
Generate a 5-second video of [DESCRIPTION_OF_SCENE] at 720p. Once generated, I want to edit it by [DESCRIBE_THE_CHANGE].
Gemini Omni Flash (gemini-omni-flash-preview) is specifically designed for high-speed video tasks. It allows for conversational editing, meaning you can ask the model to 'make this faster' or 'change the background' after an initial generation.
Source
Release notes | Gemini API | Google AI for Developers"generate 3–10 second videos at 720p from text descriptions or animate still images, and then conversationally edit and refine the outputs."