Google Releases Gemini Omni All-Round Model: Supports Multimodal Processing & Conversational Editing

At the 2026 Google I/O Developer Conference, Google officially launched the Gemini Omni model, the most comprehensive version in the Gemini family to date. Focused on "omnipotence", the model smoothly processes multimodal information like text, images, videos and audio, showing depth and flexibility.

Gemini Omni claims to "generate any output from any input" and supports conversational editing—users can modify video elements such as characters or backgrounds with just one sentence.

In addition, Google released Gemini Omni Flash, the first model of the Gemini Omni family. It is now available on Gemini App, Google Flow and YouTube Shorts, with API services coming soon.