The annual Google I/O 2026 conference took place in Mountain View, during which the company made a series of important announcements focused on the development of artificial intelligence and the Gemini ecosystem. The central themes of the event were the new Gemini 3.5 and Omni models, the AI agent Spark, an updated search, content generation services, as well as innovative devices and new pricing for artificial intelligence.
This is reported by Finway
Gemini 3.5, Omni, and Content Generation Capabilities
The standout new model was Gemini 3.5 Flash — the first model of the new generation, which has already begun to be implemented in web search AI Mode, the Gemini app, and developer tools. Gemini 3.5 Pro was also announced, which will be available in June and will be aimed at more complex computational and autonomous tasks.
“Google claims that Gemini 3.5 Flash is the most powerful model for agent tasks and programming, capable of performing some operations significantly faster than the competition.”
The updated model features reduced query costs, enhanced security, and improved answer quality. The company has implemented additional logic verification mechanisms to reduce the number of false refusals and harmful content. Gemini 3.5 is also more deeply integrated with Google’s agent tools, allowing for the automation of complex workflows, data analysis, and interface creation on the Antigravity platform. Partners are already using this model to work with large corporate datasets and research tasks.
Another important innovation was the Gemini Omni model — a multimodal AI for generating and editing videos. Omni can create videos from text, images, audio, and other video clips, and in the future — images and audio directly within the model. Users can change the environment, style, camera angle, and other video parameters, as well as create personal digital avatars with their own voice. All materials created using Omni are automatically marked with the digital watermark SynthID. Free access to Omni will be limited through YouTube Shorts and YouTube Create, while extended features will be available to subscribers of Google AI Plus, Pro, and Ultra.
AI Agent Spark, App Updates, and New Pricing
Among the biggest announcements was the personal cloud AI agent Gemini Spark, which can perform user tasks without the need to keep the device powered on. The agent integrates with Gmail, Docs, Workspace, as well as third-party services like Canva and OpenTable, and can analyze emails, gather documents, and track subscriptions. User confirmation is required for important actions.
The Gemini app has also been updated: a new Neural Expressive design has been introduced, along with an improved Gemini Live voice mode and the Daily Brief feature for generating a morning event overview based on Gmail and Calendar data.
The company has changed its approach to search by introducing an intelligent search bar based on AI Mode. The system now accepts text, images, videos, files, and Chrome tabs, better understands user intent, and offers more accurate query refinement options. AI Mode operates on the basis of Gemini 3.5 Flash, and the Personal Intelligence feature allows the use of data from Gmail, Photos, and Calendar. Agent functions have been added, enabling the AI to monitor news, blogs, real estate, and quotes in the background, generating overviews and recommendations for the user.
Google has introduced new pricing for AI services: the AI Ultra plan has been reduced to $200 per month, while the new Google AI Ultra at $100 per month provides access to advanced features of Gemini, Antigravity, and 20 TB of cloud storage along with YouTube Premium. The Antigravity 2.0 platform has become a separate direction — a service for quickly creating interfaces, mini-apps, tables, charts, and simulations using artificial intelligence.
Special attention has been paid to the integration of AI into Gmail, Docs, and Keep. New voice features allow users to search for information in emails using natural language, convert speech into structured documents, and create notes with lists during conversations.
Project Genie, Android XR, and Scientific Initiatives
The company has expanded the capabilities of Project Genie — a tool for creating interactive 3D worlds with support for Google Street View. For the first time at the conference, Android XR glasses were showcased, created in collaboration with Samsung, Gentle Monster, and Warby Parker. The devices provide voice interaction with Gemini, real-time speech translation, and camera functionality.
The final highlight was the announcement of the Gemini for Science program, which opens access for researchers to new tools for working with the UniProt, AlphaFold, and AlphaGenome databases. More than 100 scientific organizations, including Stanford University and Imperial College London, have already joined the initiative.