Google has launched two new autonomous research agents, Deep Research and Deep Research Max, built on its Gemini 3.1 Pro model. Both are now in public preview through the paid tiers of the Gemini API for developers automating heavy-duty research.
- Deep Research prioritizes speed with lower latency and cost, replacing the December preview version.
- Deep Research Max prioritizes depth using extended test-time compute for asynchronous background workflows.
- Agents support the Model Context Protocol (MCP) to pull from proprietary data streams alongside the open web.
- The agents can generate native charts and infographics in HTML or "Nano Banana" format directly within reports.
- Google claims Deep Research Max shows significant improvements on retrieval and reasoning tasks compared to its predecessor.
The agents run on infrastructure powering consumer products like Gemini app and NotebookLM, with a rollout to startups and enterprise customers planned via Google Cloud.