Google debuts Gemini Omni model

Google has unveiled Gemini Omni, its most advanced multimodal AI model to date, at the I/O 2026 conference.

The new model can generate high-fidelity video content and perform complex editing tasks through simple conversational prompts.

Gemini Omni integrates real-world physics and reasoning to ensure realistic motion and spatial consistency in its outputs.

This launch represents a major step in Google's effort to maintain its leadership in the generative AI space against rivals like OpenAI and Anthropic.