Google debuts Deep Research agents on AI Studio and APIs

Google has launched two evolved research agents, Deep Research and Deep Research Max, through its Gemini API. These agents are targeted at developers, enterprises, and startups in sectors like finance and life sciences that require rigorous, autonomous research workflows. Both agents are now available in public preview via the paid tiers of the Gemini API, with further plans to expand access to Google Cloud customers.

Deep Research is optimized for speed and direct user experience, while Deep Research Max is tailored for asynchronous, large-scale analytical tasks that require thoroughness and source diversity.

These agents leverage the Gemini 3.1 Pro model, which marks a shift from basic summarization to comprehensive, enterprise-grade analysis. The new agents can synthesize information from the open web, proprietary data via Model Context Protocol (MCP), and various file types, generating reports complete with native charts and infographics. Features like collaborative planning, multimodal input support, and real-time streaming of intermediate outputs provide developers and analysts with transparency and control over the research process.

Compared to previous versions, Deep Research Max consults more sources, resolves conflicting evidence, and produces nuanced, cited reports, setting it apart from earlier releases and comparable offerings in the market.

Google is collaborating with established data providers such as FactSet, S&P Global, and PitchBook, enabling integration of specialized financial data directly into research workflows. This expansion builds on Google’s previous work with the Gemini Deep Research agent, reaffirming the company’s focus on providing scalable, autonomous research infrastructure for professional and regulated environments.

Source