The Gemini Omni Flash model has been launched today.
At Google I/O 2026, Google introduced a new AI model named Gemini Omni, primarily showcasing its video-generation capabilities. The first in the Omni AI line, the Gemini Omni Flash, is now available.
The Omni model is touted by Google DeepMind CEO Demis Hassabis as a significant step toward AGI. Unlike traditional text-to-video tools, Omni is multimodal, handling input and output in text, audio, images, and video formats, generating an interactive world with real-world understanding.
Omni can generate videos with realistic physics, recognizing prompt contexts for accurate content. Users can edit videos through conversations with the model, altering styles, angles, scenery, or details. Gemini Omni Flash allows users to create digital likenesses, with responsible feature testing ongoing. All videos come with a SynthID watermark for AI verification.
Available now to Google AI Plus, Pro, and Ultra subscribers, Omni Flash will also roll out on YouTube Shorts and the YouTube Create app for free later this week.