Google has launched its new advanced AI research agents: Deep Research and Deep Research Max. The agents are built on the company’s latest Gemini 3.1 Pro model. They are designed to automate complex research work. In a single click, both tools can generate fully structured and sourced reports.
Google’s new Deep Research and Deep Research Max agents are here
Both the Deep Research and the Deep Research Max agents are built for specific purposes. The standard Deep Research focuses on speed and efficiency. It replaces an earlier preview build with an improved quality promise. It also claims reduced latency and costs. The agents’ low-cost, high-speed, and efficiency features make it a perfect choice for chat-based interfaces where users expect quick and reliable responses.
On the other hand, the Deep Research Max is built for stronger analysis and research work. It focuses more on deep reasoning and computation than speed. Google says the particular agent works best in non-parallel workflows. As per internal benchmarks, the Max agent outshines the standard variant in complex research and reasoning tasks. It was able to gather more sources and identify nuances that the standard model missed.
Further, Google has confirmed that both agents support the Model Context Protocol. This basically means the developers can connect their proprietary data sources, such as financial or enterprise data, to the agents. This particular feature makes the agents function beyond basic web browsing. It can be used as a fully autonomous system that’s capable of querying specialized datasets. Moreover, the agents have support for multiple input formats, including images, videos, audio, and documents.
Google has also compared it with similar tools from OpenAI and Anthropic and has highlighted its agent’s strong performance gains. However, the direct comparison remains complex, which makes the claim less reliable. Both agents’ deployment is expected in the coming days through Google Cloud.
