Google has begun rolling out an updated Gemini 2.0 Flash Thinking Experimental reasoning model that is faster and supports new features like file upload. The model, which is trained to break down prompts into a series of steps for more indepth responses, will now power Google Deep Research, replacing Gemini 1.5 Pro, and broadening its access so “anyone will be able to try it across the globe in 45-plus languages,” according to Google. Previously Deep Research was only available to subscribers paying $20-per-month for Google One AI Premium. Deep Research is a sort of AI assistant that searches the web, synthesizing information.

“For the uninitiated, Deep Research allows you to ask Gemini to create comprehensive but easy-to-read reports on complex topics,” Engadget writes, adding that it “works slower than your typical chatbot, and that’s by design.”

VentureBeat notes that while 2.0 Flash Thinking Experimental is less powerful than the Gemini Pro 1.5 subscription offering, it found it nonetheless “a helpful research assistant.”

Google explains in a blog post that the Thinking model “enhances Gemini’s capabilities across all research stages — from planning and searching to reasoning, analyzing and reporting — creating higher-quality, multi-page reports that are more detailed and insightful.”

“Gemini 2.0 Flash Thinking Experimental will also power a new capability called personalization,” reports VentureBeat, explaining that “responses will be more tailored to the user by referencing previous conversations or searches.”

To provide this heightened personalization, Gemini 2.0 Flash has been connecting to Google apps and services, including Calendar, Notes, Search, Photos and YouTube. Google says the AI will only use such information with permission.

“You can ask Gemini for restaurant recommendations and it will reference your recent food-related searches,” according to a Google blog post. Personalization must be enabled, and can also be disabled. Users are directed to Gemini’s personalization page for more information.

“Think of it as RAG, but for a Gemini user rather than an entire organization,” reports VentureBeat, referring to Retrieval Augmented Generation, a method of improving the accuracy and reliability of generative AI with information collected from focused, relevant data sources.

Another feature rolling out “for everyone at no cost in the Gemini app,” Google says, is Gems. “Gems let you customize Gemini to create your own personal AI expert on any topic. Get started with one of our premade Gems or quickly create your own custom Gems, like a translator, meal planner or math coach,” the company suggests.

ZDNet believes the now-free Gems and Deep Research are “two of Gemini’s best features,” adding that the updates have taken Google “from being a dark horse in the AI space to a frontrunner.”