Google introduces Gemini Omni Flash for video generation and editing

Google announced the first Gemini Omni model at I/O 2026. It accepts text, images, audio, and video and uses knowledge of physics, science, and history to generate and edit video from natural-language prompts. The Omni architecture combines generative media models with Gemini's knowledge. Google planned to start with products such as YouTube Shorts, followed by Gemini API and Agent Platform API access.

Visit original source