Just this month, Google introduced the Deep Research feature in Gemini for Android, which is now being expanded to the iPhone app for Gemini Advanced subscribers.
Deep Research “utilizes advanced reasoning and extensive contextual understanding to function as a research assistant, delving into intricate subjects and compounding reports on your behalf.” This marks Gemini’s first “agentic” capability and was officially introduced at gemini.google.com in December.
You initiate the process by posing a research question that can be as lengthy and detailed as necessary, incorporating multiple elements. Gemini will then provide a step-by-step plan that you can modify or remove sections from.
Upon clicking “Start research,” Gemini typically takes 5-10 minutes to scour the web, examine various sites, and conduct additional searches based on its findings. This cycle continues multiple times.
More intricate reports will require additional time to finalize, and you are free to exit the app once research commences. The Gemini app will notify you once the process has concluded, or you can revisit the Chat history to access the information later.
You will receive a report arranged into distinct sections with tables (as applicable) and sources cited at the end. There’s also a handy option to export the report to Google Docs.
Moreover, there are constraints regarding “Daily research requests” and the “Number of simultaneous research requests” you can run. The Gemini app will keep you informed about the remaining research requests for the day.
Currently, Deep Research is powered by Gemini 1.5 Pro. It is expected that Google will transition to using 2.0 Pro once it moves out of the experimental stage, though the timeline for this upgrade remains uncertain.
Within the model selector at the top of the iPhone app, “1.5 Pro with Deep Research” can be found alongside Gemini 2.0 Flash, the new 2.0 experimental models, and the soon-to-be-retired 1.5 family.
FTC: We use income earning auto affiliate links. More.
