Gemini Advanced Gets "Deep Research" Feature
Gemini Advanced users now have access to a new "Deep Research" capability, marking the first Gemini feature to embody Google's vision of agents performing complex actions. This feature was initially previewed at Made by Google 2024.
How Deep Research Works
Users pose research questions to Gemini, which then creates a multi-step plan. This plan can be revised and refined before execution. Once started, Gemini searches the web, iteratively refining its analysis based on the information gathered. The process involves multiple searches, with each new search building upon previous findings.
Key Features and Benefits
- Comprehensive Reports: Gemini generates organized reports with key findings, sources, related content, and links to researched websites.
- Time-Saving: This "personal research assistant" completes research in minutes, compared to hours of manual work.
- Easy Exporting: Reports can be exported to Google Docs for further use.
Availability
Deep Research is currently available for Gemini Advanced users in English on desktop and mobile web. Mobile app support is expected in early 2025. Users can access the feature by selecting "Use 1.5 Pro with Deep Research" in the model picker. For more on Gemini's development, see Gemini UI Refresh and Google's Willow project. For related AI news, check out this article on AI-generated voices.