Google Launches Gemini 3.1 Pro-Based Deep Research Agents, Connecting Proprietary Data Sources to Generate Professional Charts

Google recently posted a blog announcing the upgrade of its autonomous research agents, launching Deep Research and Deep Research Max based on the Gemini 3.1 Pro model. The former is suitable for interactive scenarios, balancing speed and efficiency; the latter is for asynchronous background workflows, pursuing comprehensiveness and high quality through extended computing for iterative reasoning.

The core breakthroughs of these two agents lie in expanded data sources and output forms: they can search the web, remote MCP servers, upload files, and connect to storage. Users can securely link to professional data streams like finance or market data. For output, they can natively generate high-quality charts and infographics for dynamic visualization of complex data.

New features also include collaborative planning, extended tool sets, multimodal foundations, and real-time streaming output. Users can review and optimize research plans before execution, combining tools like Google Search, MCP servers, and code execution.