Google has unveiled significant enhancements to its Gemini app, making several premium features more accessible while introducing new capabilities that promise a more personalized AI assistant experience.
Summary
Google has announced substantial upgrades to the Gemini app, improving both performance and availability of its most popular features. The company is rolling out an upgraded version of its 2.0 Flash Thinking Experimental model with added capabilities like file upload, improved efficiency, and speed. For Gemini Advanced users, this model now offers an expanded 1M token context window, enabling analysis of larger information sets for complex problem-solving.
The Deep Research feature, initially introduced in December, has received a significant boost by leveraging the 2.0 Flash Thinking Experimental model. This enhancement improves all research stages—from planning and searching to reasoning, analyzing, and reporting—resulting in higher-quality, multi-page reports. A notable improvement is the ability to observe Gemini’s thought process in real-time as it browses the web. This feature is now rolling out globally in over 45 languages, with free users getting limited monthly access and Advanced users receiving expanded capabilities.
Google is also introducing a new experimental personalization feature that connects Gemini with Google apps and services, beginning with Search. This integration allows for more tailored responses based on users’ search history, such as restaurant recommendations informed by past food-related searches. Users maintain control over their data and can disconnect Search history from Gemini after enabling the feature.
Connected apps integration has been enhanced with Calendar, Notes, Tasks, and Photos now available on the 2.0 Flash Thinking Experimental model. This improvement enables Gemini to better handle complex multi-app requests. In the coming weeks, Google Photos integration will allow users to leverage their personal photos for various tasks, such as creating travel itineraries based on trip photos.
Finally, Gems—a feature that allows users to create customized AI experts on specific topics—is now available to everyone at no cost. Users can either select from premade Gems or create custom ones through the “Gems manager” on desktop, with the added ability to upload files for enhanced functionality.
Key Takeaways
- Upgraded 2.0 Flash Thinking Experimental model offers file upload support and a massive 1M token context window for Advanced users.
- Deep Research has been enhanced with the 2.0 Flash Thinking Experimental model and is now available globally in 45+ languages.
- New personalization feature connects with Google Search to provide more tailored responses based on search history.
- Connected apps integration now works with Calendar, Notes, Tasks, and Photos on the 2.0 Flash Thinking Experimental model.
- Google Photos integration is coming soon, allowing users to leverage their personal photos for various AI assistance tasks.
- Gems customization feature is now available to all users at no cost, enabling creation of personalized AI experts.
- Free users can access premium features like Deep Research a limited number of times per month.
Links
Announcement: New gemini app features march 2025
Demo live: https://gemini.google.com/deepresearch?hl=en
