Gemini Omni merges Gemini’s reasoning capabilities with creative potential, while Gemini 3.5 is designed to handle complex tasks and workflows. These models were unveiled at Google I/O 2026, featuring the latest innovations: Gemini Omni and the Gemini 3.5 model series.
The Gemini Omni model allows users to create diverse outputs from various inputs, including video, images, audio, and text. This model leverages Gemini’s knowledge of the real world to generate high-quality videos, which users can edit simply through conversation.
Additionally, Gemini 3.5 represents a significant advancement in creating intelligent agents. The series begins with the release of 3.5 Flash, which is optimized for delivering high-performance results in agentic and coding tasks, particularly for tasks that require a long-term approach and practical application.
Interactive Video Editing
One of Omni’s standout features is its ability to facilitate video editing through natural language instructions. Each command builds on the previous one, maintaining character consistency and realistic physics, allowing users to significantly alter their video scenes.
For instance, users can prompt Omni to change the lighting or introduce new elements into a scene. A specific example involves creating a video loop where a black and white checkerboard room is nested within a glass sphere, which is held above a hand.
Complex Agentic Tasks
3.5 Flash, when powered by Antigravity, excels at executing lengthy agentic tasks efficiently. It supports multi-step workflows and coding, making it suitable for categorizing and renaming unstructured assets dynamically.
The combination of 3.5 Flash with the Antigravity harness enables it to deploy subagents for large-scale problem-solving. It can generate various creative outputs, such as visualizations of mathematical concepts and numerous fractal variations at high speeds.
Innovation in Search and AI Integration
Gemini 3.5 Flash is integrated into the Gemini app and AI Mode in Search globally. This integration introduces new intelligent features, such as information agents that provide timely updates and insights, launching first for Google AI Pro & Ultra subscribers.
Search now leverages 3.5 Flash to create custom generative UIs tailored to user queries, offering visual tools and simulations. These features will be accessible to all users this summer.
Furthermore, Search will assist with ongoing projects by developing custom dashboards and trackers, initially available to Google AI Pro and Ultra subscribers in the U.S.
Introducing Gemini Spark
Gemini Spark, a personal AI agent, operates on Gemini 3.5 with the Antigravity harness. Available to Google AI Ultra subscribers in the U.S., it integrates with Workspace tools, assisting users in their digital activities.
Gemini Omni Flash is being made available globally to Google AI Plus, Pro, and Ultra subscribers, and will soon be accessible to developers and enterprises through APIs.