Google DeepMind's Deep Research Max runs 160 searches overnight for exhaustive reports

Google DeepMind has launched two new autonomous research agents—Deep Research and Deep Research Max—now in public preview via the Gemini API. Built on Gemini 3.1 Pro, the standard Deep Research optimizes for speed and cost, ideal for interactive products. Deep Research Max targets asynchronous overnight workflows, running around 160 search queries per task using extended test-time compute to iteratively refine output. Benchmarks show 93.3% on DeepSearchQA and 85.9% on BrowseComp. The agents search the open web, user file uploads, and connected data sources through Model Context Protocol (MCP) servers, generating native charts and infographics while consulting over 100 sources per task. Google is partnering with FactSet, S&P Global, and PitchBook for dedicated MCP integrations to plug financial data feeds directly into workflows.

The agents accept PDFs, CSVs, images, audio, and video as inputs, and support collaborative planning where users review and edit the research plan before execution. Real-time streaming surfaces reasoning steps as they happen. Google positions the launch as a move beyond summarization into an enterprise workflow layer for finance, life sciences, and market research. The technology already powers features in Gemini App, NotebookLM, Google Search, and Google Finance. Neil Hoyne, Google's Chief Strategist, emphasized: 'No. Unfortunately, this won't replace good judgment… But for anyone tired of drowning in tabs, half-read PDFs, maybe this kind of AI gives you your evenings back.' The agents put Google in direct competition with OpenAI and Anthropic on agentic research tooling, with implications for education and EdTech teams using AI for research and curriculum analysis.