Google is placing artificial intelligence at the core of its strategy, introducing new Gemini models and agentic capabilities designed to integrate deeply into its most used services. The company unveiled Gemini 3.5 Flash, a faster iteration of its AI model, and Gemini Omni, a multimodal model capable of generating video from various inputs. These advancements are set to power a revamped Google Search experience, which will feature AI agents capable of performing complex research and tasks. Additionally, Google previewed new "intelligent eyewear," developed in partnership with Samsung and fashion brands, expected to ship this fall. This broad push indicates Google's intent to embed AI agents across its entire ecosystem, from Search and Workspace to its hardware offerings.
Google announced a suite of AI-driven updates at its I/O 2026 developer conference, emphasizing the integration of its Gemini models into core products. The company unveiled Gemini 3.5 Flash, a new model designed for speed and efficiency, suitable for agentic tasks and coding. This model is now powering the Gemini app and AI Mode in Google Search globally. Alongside Gemini 3.5 Flash, Google introduced Gemini Omni, a multimodal model capable of generating video from text, images, and video prompts. Google DeepMind CEO Demis Hassabis described Gemini Omni as a step toward artificial general intelligence.
The company also detailed significant upgrades to Google Search, positioning it as an AI-first product. The search box has been redesigned to accept a wider range of inputs, including text, images, files, and video, and will dynamically expand to accommodate more natural language queries. This update, described as the biggest change to the search box in over 25 years, is intended to move beyond keyword-based searches to more conversational interactions. A key feature of the new Search is the introduction of "information agents," autonomous AI bots that can perform in-depth research, monitor the web for updates, and provide summaries or action plans. These agents are designed to operate 24/7 in the background, providing proactive alerts for users. For subscribers to its AI Pro and Ultra plans, Search will offer the ability to create these agents and generative user interfaces that can build custom visuals and dashboards.
Google also announced Gemini Spark, a new AI agent designed to provide proactive, 24/7 assistance. Gemini Spark can access information from users' Gmail, Google Calendar, and other Google products to perform research, plan trips, and manage tasks. It will be available to Google AI Ultra subscribers in the U.S. next week and will integrate with Chrome later this summer. The company is also making advancements in its Workspace suite, with features like "Docs Live" allowing users to speak their thoughts and have Gemini convert them into structured documents in real time.
In hardware, Google previewed new "intelligent eyewear" set to launch this fall. Developed in partnership with Samsung, and featuring frames from Warby Parker and Gentle Monster, these glasses will respond to voice commands via Gemini for tasks such as navigation and communication management. The glasses will also include embedded cameras for taking photos and videos. This move into smart glasses follows Google's earlier attempt with Google Glass and competes with similar offerings from companies like Meta.
The company also introduced Gemini 3.5 Flash, a model optimized for speed and efficiency, which is now the default for the Gemini app and Google Search's AI Mode. Gemini 3.5 Flash reportedly outperforms Gemini 3.1 Pro on several benchmarks and is significantly faster. For developers, Google launched Managed Agents in the Gemini API, enabling the creation of autonomous agents within secure sandbox environments. The Antigravity Agent, a general-purpose managed agent capable of planning, reasoning, writing code, and browsing the web, was also released in public preview.
Google also announced an AI Ultra tier at $100 per month, aimed at developers, creators, and power users, which includes access to Gemini Spark. The company is also enhancing AI content identification, with C2PA content credentials coming to Gemini and Chrome to help users determine if images were captured by a camera or generated/edited by AI.
The announcements signal Google's commitment to embedding AI agents across its product line, aiming to transform user interaction with its services. Google CEO Sundar Pichai stated that the company is in an "agentic Gemini era," highlighting the focus on AI agents that can proactively perform tasks for users.