Google just gave Gemini a major upgrade – Here’s what’s new
Google has unveiled a significant upgrade to its Gemini AI platform, introducing a suite of new features and enhancements designed to make the AI more powerful, personal, and accessible. The updates focus on improving reasoning, research, and personalization, solidifying Gemini's position as a leading AI assistant.
“Flash Thinking Experimental” gets smarter and more efficient
A key highlight is the upgraded 2.0 Flash Thinking Experimental model, now boasting a 1 million token context window. This expanded capacity enables users to tackle complex tasks involving vast amounts of information. Moreover, the model now supports file uploads, enhancing its ability to analyze and process diverse data types.
“Deep Research” rolls out globally
Google has also significantly improved Deep Research, now powered by the 2.0 Flash Thinking Experimental model. This upgrade elevates Deep Research's capabilities across all stages, from planning and searching to analysis and reporting. Users can expect higher-quality, multi-page reports with deeper insights.
Notably, Gemini now displays its reasoning process in real-time while browsing the web, offering transparency into its research methodology.
"We've seen incredible enthusiasm about Deep Research, and are rolling it out starting today so that anyone will be able to try it across the globe in 45+ languages," Google stated in its blog post. Deep Research is now accessible to all Gemini users, with Gemini Advanced subscribers receiving expanded access for more demanding projects.
Increased integrations for better personalization
A major push towards personalization is evident in the introduction of an experimental feature called personalization, also powered by the 2.0 Flash Thinking Experimental model. By connecting with users' Google Search history, Gemini can deliver more tailored responses.
For instance, it can provide restaurant recommendations based on recent food searches or travel advice based on previously researched destinations. Users retain full control over their Search history and can disconnect it from Gemini at any time.
The integration of Google apps and services has also been enhanced. The 2.0 Flash Thinking Experimental model now supports additional connected apps, including Calendar, Notes, and Tasks. This allows Gemini to handle complex, multi-app requests more effectively.
For example, users can ask Gemini to find a recipe on YouTube, add the ingredients to a shopping list, and locate nearby open grocery stores, all in a single prompt. Google also announced that Google Photos integration is coming soon, enabling users to leverage their photo library for tasks like creating travel itineraries or retrieving information.
Gems now available to all users
Finally, Gems, the customizable AI experts, are now available to all Gemini users at no cost. Users can create personalized Gems for various tasks, such as translation, meal planning, or math tutoring. The ability to upload files during Gem creation further expands their utility.
These updates represent a significant leap forward for Gemini, reinforcing Google's commitment to developing AI that is both powerful and personalized. The enhanced reasoning capabilities, improved research tools, and deeper integration with Google services promise to make Gemini an even more indispensable assistant for users worldwide.
