As reported by 9to5Google, users of Gemini Advanced now have access to the Deep Research feature via the Gemini 2.5 Pro (experimental) model. Previously, Deep Research was available to all users but was based on the Gemini 2.0 Flash Thinking model.
This recent update allows Gemini users to utilize Google’s most advanced model for deep research tasks, enhancing the quality of the reports generated. For those unfamiliar with the concept, Deep Research is designed to streamline the research process.
Users provide a topic, and Gemini formulates a research plan, scours hundreds of websites, and produces a detailed report tailored to the specified needs. With the integration of Gemini 2.5 Pro, even in its experimental state, Google asserts that the reports generated are more precise and insightful than before.
Google states that “raters preferred the reports generated by Gemini Deep Research powered by 2.5 Pro over other leading deep research providers by more than a 2-to-1 margin.” In comparison to OpenAI’s deep research model, Gemini 2.5 Pro (experimental) has shown a preference rate of 69.9%.
It also excelled in various criteria, including instruction adherence (60.6% vs. 39.4%), comprehensiveness (76.9% vs. 23.1%), completeness (73.3% vs. 26.7%), and writing quality (58.2% vs. 41.8%). To access the Deep Research feature with Gemini 2.5 Pro (experimental), users can visit gemini.google.com and choose the Deep Research option, or use the Gemini app available on Android and iOS.
However, it is important to note that this feature is currently restricted to paid users, although it is possible that Google will eventually make it available to free users.
Leave a Reply