Google Launches Gemini 3.5 and Omni: Agents, Video Generation, and New Hardware

In May 2026, Google announced its entry into the era of agentic AI models. The company introduced Gemini 3.5—a model with powerful capabilities for solving multi-step tasks—and Gemini Omni, a new generation of AI capable of generating high-quality video content from combinations of text, audio, video, and images.

Gemini 3.5: The Agent That Acts

Gemini 3.5 is positioned as "intelligence plus action at the frontier"—a model that doesn't just analyze information but executes complex multi-step workflows. This means AI can independently set sub-goals, select tools, and make decisions while performing a task.

Gemini Omni expands what's possible in video generation. The model accepts combinations of text, audio, video, and images as input, creating high-quality video content from them. For content creators and multimedia companies, this opens new possibilities for automation and creativity.

From Laptops to Trackers

Google announced new hardware specifically designed to work with agentic AI:

• Googlebook—a next-generation laptop designed from the ground up for Gemini integration. The device features Magic Pointer—a function providing contextual hints and synchronization with the Android ecosystem.
• Fitbit Air—Google's most compact fitness tracker. It monitors heart rate 24/7, detects atrial fibrillation, tracks blood oxygen levels, and monitors sleep phases.
• Google Health App—an application that aggregates all health data in one place, providing a comprehensive overview of physical well-being.

These devices function as tools for interacting with AI agents—they provide agents with context about the user and their environment.

AI Across All Google Services

The company is embedding agentic capabilities across its entire ecosystem. Search received its biggest update in 25+ years: AI agents now monitor online content 24/7 and deliver updates to users with actionable links.

The Gemini app has been redesigned as a "proactive assistant"—the application independently reads emails, creates meeting schedules, and offers personalized daily briefings through the new Gemini Spark tool.

Universal Cart was introduced—a unified shopping hub integrated into Google Search, Gemini, YouTube, and Gmail. Additionally, Android Halo launched—a new interface for managing AI agents and receiving contextual assistance.

"This is a shift from AI that advises to AI that acts,"

Google positions its new era of development.

Science and Creativity Through AI

Google launched several initiatives aimed at applying agentic AI to science and creativity. Gemini for Science expands scientific research capabilities, while AlphaEvolve helps optimize supply chains and molecular systems. The REPLIQA program allocates $10 million to five universities in quantum computing and AI in biology and medicine.

A partnership with Believe equips musicians and producers with Google Flow Music—an AI assistant that aids at every stage of music creation.

What This Means

Agentic AI is moving from laboratories into real products. If AI previously served as an advisor, it now becomes a colleague handling routine and complex multi-step processes. For users, this means having personal AI assistants for everything—from email management to content creation. For companies, this will reshape user interactions: instead of offering advice, AI will propose concrete actions.