Google Rolls Out Deep Research Feature For Gemini 2.5 Pro Experimental Users; Details Here

Google has officially introduced Deep Research, a new feature now available to Gemini Advanced subscribers. The company says it’s the most capable artificial intelligence system from its stable so far.

Deep Research is now available on Gemini 2.5 Pro Experimental, allowing users to explore complex topics in greater detail with less manual effort. The model is designed to handle layered and advanced tasks and has shown strong performance in AI benchmarking platforms such as Chatbot Arena and LMArena, which use human ratings to assess AI capabilities.

Gemini 2.5 Pro also performs well in technical evaluations, including tests focused on coding, mathematics and scientific reasoning.

With the integration of Deep Research, Gemini 2.5 Pro now offers added utility for researchers, professionals, students and others seeking time-saving tools for gathering information. 

Deep Research, according to Google, is capable of being a research tool rather than just a basic AI assistant. When a user submits a query, the system creates a step-by-step guide to investigate the topic in depth.

Google reports that internal testing shows a strong preference for research outputs from this model. In fact, users rated Deep Research on Gemini 2.5 Pro as superior to competing tools by more than a 2-to-1 margin, citing improvements in clarity, depth of analysis, and the ability to synthesise diverse information into coherent conclusions.

Deep Research is currently available through Google AI Studio and the Gemini mobile app on Android and iOS for Gemini Advanced users. It will soon be rolled out to Vertex AI, Google's cloud-based AI platform for enterprise use. 

Google has also announced plans to introduce pricing tiers in the near future, allowing businesses and power users to scale their usage with higher rate limits suitable for production environments.

In addition to generating written reports, the tool offers an Audio Overviews feature — a unique function that converts the content into a podcast-style audio format. This enables users to consume the information hands-free while multitasking, travelling or just taking a break from screen time.

To try it out, users can open the Gemini interface, choose the “Gemini 2.5 Pro (Experimental)” option from the dropdown, and click on the “Deep Research” button in the prompt bar to begin generating reports.