Google has officially debuted its new autonomous research agents – Deep Research and Deep Research Max – powered by Gemini 3.1 Pro. This update is now accessible in preview via the Gemini API.
Deep Research Max handles complex, long – term tasks. It uses extended compute to refine results and build detailed reports.
The standard Deep Research version focuses on high speed. It replaces the previous December model to work better with various low – latency systems.
Agents access corporate data via Model Context Protocol. This connects internal databases, financial, and analytical platforms.
The system supports visualization. It generates charts using HTML or Nano Banana directly in the output.
Users can mix files, images, video, and audio. A tracker monitors processing steps in real – time.
Google notes that Max uses more sources and manages contradictory data better. This provides detailed reports that account for a variety of different perspectives.
Targeted at finance, science, and analytics, Google is partnering with providers to integrate these tools into workflows.
