Gemini Omni is Googles new world model, with advanced AI video generation capabilities
Source: Mashable Tech
Overview
Google unveiled a new AI world model at Google I/O 2026 called Gemini Omni. The company describes Gemini Omni as “a new model that can create anything from any output,” and the showcase highlighted its advanced video‑generation capabilities. The first release in the Omni family is Gemini Omni Flash.
Credit: Google
Gemini Omni Flash vs. Other AI Video Tools
Gemini Omni Flash differentiates itself from existing tools such as Google Veo and the earlier Genie world model (Google’s Genie world model). During the I/O presentation, DeepMind CEO Demis Hassabis called Omni a crucial step toward AGI, emphasizing that future versions will be able to output “anything” a user requests.
Multi‑modal Input & Output
Unlike many text‑to‑video systems, Gemini Omni is multi‑modal for both input and output:
- Inputs: text, audio, images, video
- Outputs: generated video, interactive worlds, and other media
The model leverages Gemini’s extensive real‑world knowledge to create unique, context‑aware content.
Video Generation Quality
Google claims Omni can produce videos with:
- More accurate physics, resulting in realistic‑looking scenes
- Contextual understanding (e.g., historical facts) to improve factual accuracy
Credit: Google
Conversational Video Editing
Omni also supports editing through conversation. Users can:
- Modify background, style, angle, scenery, or specific details of an existing or AI‑generated clip
- Interactively refine videos by describing desired changes
Credit: Google
Avatars & SynthID Watermark
- Avatars: Gemini Omni Flash can create a digital version of a user’s likeness, though this feature is still in testing to ensure responsible launch.
- SynthID watermark: All generated videos are embedded with a SynthID watermark, allowing verification that the content was AI‑created.
Availability
- Today: Gemini Omni Flash is rolling out to paid Google AI Plus, Pro, and Ultra subscribers via the Gemini app and Google Flow.
- Later this week: The feature will launch in YouTube Shorts and the YouTube Create app at no cost to users.