Google launched two new autonomous research agents, Deep Research and Deep Research Max, allowing developers to integrate open web data with proprietary enterprise information through a single API call. This update is the most significant since the product’s debut and aims to bolster Google’s presence in enterprise research workflows for industries like finance and life sciences, where accurate information is critical.
Built on Google’s Gemini 3.1 Pro model, the release positions these agents as tools that can streamline exhaustive multi-source research, which has traditionally required significant human analyst time. “We are launching two powerful updates to Deep Research in the Gemini API,” said CEO Sundar Pichai on X. “Use Deep Research when you want speed and efficiency, and use Max when you want the highest quality context gathering & synthesis using extended test-time compute.”
Deep Research targets low-latency, interactive use cases, offering reduced latency and cost while maintaining high quality. Conversely, Deep Research Max is designed for more thorough context gathering, relying on extended computational cycles to generate detailed analyses. This tiered architecture reflects the balance between speed and thoroughness in AI agent design.
The newly introduced Model Context Protocol (MCP) support allows both agents to query secure private databases and third-party data services. Google is collaborating with FactSet, S&P, and PitchBook on MCP server designs to seamlessly integrate financial data into workflows reliant on Deep Research. This initiative aims to address the gap between open internet information and the specific data organizations need for decision-making.
Deep Research’s capacity for multimodal input and its ability to generate native charts and infographics directly in reports enhance its utility. The inclusion of visual elements eliminates the need for users to export data for graphics, improving the tool’s functionality for professional environments. AI commentator Shruti Mishra highlighted this as “actual rendered charts inside the markdown output.”
Deep Research was initially introduced as a consumer feature in December 2024, evolving through multiple iterations, including upgrades that significantly improved its benchmark performance. The release positions Google as a strong competitor in the autonomous research agent market, which is also being pursued by rivals like OpenAI and various startups.
The pricing structure for Deep Research is reportedly competitive, further encouraging developer adoption. Issues have been raised regarding the exclusivity of the new agents to the API, as they are not available in the Gemini consumer app, prompting concerns among users about equitable access for app subscribers.
Deep Research and Deep Research Max are currently in public preview through paid tiers of the Gemini API, with future availability expected on Google Cloud for startups and enterprises. The success of these agents will largely depend on their ability to meet professional standards for quality and reliability in various sectors. “This is just the start of our agents journey,” stated Logan Kilpatrick, who leads developer relations for Google’s AI efforts.
