Google launched two autonomous research agents built on Gemini 3.1 Pro—Deep Research and Deep Research Max—in public preview via paid Gemini API tiers on April 21. Both support Model Context Protocol connections to private and commercial data sources (FactSet, S&P Global, PitchBook) and natively generate charts and infographics; Deep Research Max uses extended compute for comprehensive asynchronous research and scores 93.3% on DeepSearchQA (up from 66.1% in December) and 54.6% on Humanity's Last Exam (up from 46.4%). Full API documentation is available.