Google Unveils Gemini 3.5, Omni Model, and Agentic Tools at I/O 2026
Google introduces a suite of new AI models and developer tools at I/O 2026, headlined by the Gemini 3.5 family and a multimodal Omni model. These innovations aim to advance the Agentic Enterprise directly through Google Cloud and Workspace platforms.
Google reveals a groundbreaking lineup of AI innovations at I/O 2026 designed specifically for Google Cloud customers. The headliners include the Gemini 3.5 family of models, which kicks off with the 3.5 Flash version that delivers flagship-level intelligence specifically optimized for agentic workflows and complex coding tasks. Built from the ground up using Google's custom AI infrastructure, this model rivals larger flagship models in performance while maintaining the high speeds users expect from the Flash series.
In addition to Gemini 3.5, Google introduces Gemini Omni, a major leap forward in world understanding and multimodality that allows users to generate any output from any input, starting with video. The company also debuts Gemini Spark, a 24/7 personal AI agent for Enterprise and Workspace customers that autonomously takes action under user direction. Furthermore, new Workspace tools like Google Pics for image generation and advanced voice features in Gmail, Docs, and Keep completely reimagine daily productivity tasks.
For developers and enterprise security teams, Google expands its Agent Platform with the new Managed Agents API for building and running custom agents in secure, Google-hosted environments. The platform also welcomes CodeMender, a specialized AI security agent that actively finds and fixes vulnerabilities in code. Alongside the expanded Google Antigravity capabilities, these tools provide organizations with the comprehensive infrastructure needed to fully embrace the Agentic Enterprise.