
AI · Enterprise Software
Google Launches Deep Research Agents on Gemini 3.1 Pro
Google introduced Deep Research and Deep Research Max, AI agents powered by Gemini 3.1 Pro for autonomous long-horizon research. Available now in public preview via Gemini API for developers and enterprises.
Google released Deep Research and Deep Research Max on April 21, 2026, as advanced evolutions of its Gemini Deep Research agent. These agents, built on Gemini 3.1 Pro with a 1 million token context window, autonomously handle multi-step research tasks including planning, searching across the web or custom sources, analysis, verification, and generation of detailed, cited reports with native visualizations. Deep Research prioritizes speed and efficiency for interactive user interfaces, while Deep Research Max employs extended compute for comprehensive synthesis in asynchronous workflows like due diligence or market analysis. Key features include collaborative planning for scope control, integration with tools such as Google Search, MCP servers, code execution, and file search, plus support for proprietary data streams.