Google unveils new model "Gemini Omni" for creating all kinds of content

Google announced Gemini Omni at I/O 2026, a new multimodal model for generating and editing media across text, images, audio, and video.

The first version, Gemini Omni Flash, is rolling out to the Gemini app, Google Flow, and YouTube tools, with API access coming soon.

It can keep subjects and styles consistent across prompts, edit existing footage, and add or replace characters and backgrounds.

Google also built in SynthID for AI content identification, strengthening provenance safeguards for generated media.