Google has recently rolled out two autonomous research agents, Deep Research and Deep Research Max, both built upon the robust Gemini3.1 Pro architecture. These cutting-edge tools are now accessible for preview via the premium tiers of the Gemini API. Designed to streamline and automate intricate research workflows, they propel AI agents into a realm of 'long-term computation'. The standard iteration of Deep Research is tailored for real-time conversational contexts, whereas Deep Research Max is optimized for asynchronous background operations. The latest version heralds support for the MCP (Modular Computing Protocol), endowing agents with inherent visualization prowess. Performance benchmark evaluations reveal marked enhancements in Deep Research Max, though comparisons with rival offerings should be approached with discernment. Moreover, this suite of agents boasts enhanced collaborative planning functionalities, accommodates multimodal input, and offers the flexibility to disable network connectivity to safeguard data integrity. Looking ahead, Google envisions deploying these two innovative tools in the enterprise sector through Google Cloud, steering AI towards becoming adept in-depth analysis specialists within the professional analytics landscape.