Google DeepMind Launches Deep Research Agents for Autonomous Reporting and Data Analysis

Google DeepMindGoogle DeepMind

· Updated

Google DeepMind launched Deep Research and Deep Research Max, two autonomous agents powered by Gemini 3.1 Pro that navigate the web and private datasets to generate cited reports. These agents move beyond simple chat by using extended reasoning and arbitrary MCP support to synthesize complex information into professional-grade documents with native visualizations.

Google DeepMind introduced Deep Research and Deep Research Max, autonomous agents built on Gemini 3.1 Pro that perform multi-step analytical tasks. These agents browse the web or private data to produce cited reports. This release includes support for the Model Context Protocol (MCP), a standard for connecting AI to external data.
Underlying model
Gemini 3.1 Pro
Performance tiers
Deep Research, Deep Research Max
Connectivity protocol
Model Context Protocol (MCP)
Output formats
Cited reports, native visualizations
Data sources
Web, internal docs, financial data
Availability
Gemini API

This update shifts focus to test-time compute (allocating more processing power during inference to improve reasoning). By offering a "Max" tier, the team addresses the reliability gap in agentic workloads requiring deep context gathering. Native visualizations also bridge the gap between raw data and presentation-ready outputs.

Integrate these agents via the Gemini API to automate market research or internal knowledge synthesis. Deep Research is optimized for speed, while Deep Research Max handles complex, long-running background research. Both tiers are available for developers to build custom research tools that leverage private enterprise data.

View the full update on blog.google
Google DeepMind
Google DeepMind
@GoogleDeepMind
X

Deep Research and Deep Research Max are our latest autonomous research agents powered by Gemini 3.1 Pro. They can safely navigate both the web and your custom data, like internal docs and specialized financial information, to create professional-grade, fully cited reports. 🧵

127retweets1.3klikes
View on X

Still wondering? A few quick answers below.

Google Deep Research is a set of autonomous agents powered by the Gemini 3.1 Pro model. Unlike standard chatbots, these agents can independently navigate the web and private datasets to perform multi-step research tasks. They synthesize their findings into professional-grade reports that include full citations and native data visualizations to support their conclusions.

Deep Research is optimized for speed and efficiency, making it suitable for interactive applications that require faster responses. Deep Research Max is designed for exhaustive tasks, using extra time to perform deeper reasoning and more thorough context gathering. This makes the Max version ideal for complex research projects that can run in the background.

These agents now include support for the Model Context Protocol, which is an open standard for connecting AI models to external tools and data sources. This allows the agents to securely access and analyze custom information, such as internal company documents or specialized financial data, alongside information gathered from the open web.

The Deep Research and Deep Research Max agents are available to developers through the Gemini API. This allows teams to build custom research tools and applications that leverage Google DeepMind's autonomous agent capabilities. Users can choose between the two tiers based on whether their specific use case requires rapid responses or exhaustive analytical depth.

Yes, this update introduces the ability for the research agents to natively generate presentation-ready visuals. While previous versions focused primarily on text-based synthesis, these new agents can create charts and graphs that bring data to life directly within their reports, helping users visualize the information gathered during the autonomous research process.

Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →

Google

See all AI news & updates from Google →

Share this update