Introducing Gemini Omni

Source: Google Date: 2026-05-19 URL: https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-omni/

Summary

Gemini Omni Flash is Google’s new video generation model, not a language model. It creates and edits videos from any combination of inputs (images, audio, video, text) with conversational language editing. Features physics understanding (gravity, kinetics, fluid dynamics), knowledge integration, and digital avatar creation. Includes SynthID watermarks verifiable through Gemini app, Chrome, and Google Search. Available globally to AI Plus, Pro, and Ultra subscribers; rolling out to YouTube Shorts and YouTube Create at no cost.

Implications

• Feeds the Claude Design competition thread: Google now has a creative generation surface (video) that Anthropic lacks. Claude Design does static design → code; Gemini Omni does conversational video.
• Feeds the SynthID / content provenance thread: watermarking as a first-class feature across generation surfaces positions Google ahead on AI content identification.
• The model is consumer-facing, not developer-facing — no API announced yet. Enterprise and developer access “coming soon.”