Gemini began as a collaborative effort between research and engineering to build a multimodal, scalable AI foundation model. Today, Gemini 2.5 is our most intelligent model series to date.
GENERATIVE AI
-
Google Project Astra Enhanced with Voice Control and Memory
By
–
One year after introducing Project Astra we've:
— Google AI (@GoogleAI) 20 mai 2025
🗣️Upgraded voice output to be more natural
🧠 Improved memory capabilities
💻 Added computer control
Soon, we’ll bring these capabilities to Gemini Live, Search, glasses, & more! https://t.co/XzwgFyhaUk pic.twitter.com/5Z4KRJwQ3ZOne year after introducing Project Astra we've:
Upgraded voice output to be more natural Improved memory capabilities Added computer control
Soon, we’ll bring these capabilities to Gemini Live, Search, glasses, & more! http://
deepmind.google/technologies/p
roject-astra/
… -
Google Introduces Gemini World, Universal AI Assistant for AGI
By
–
BREAKING 🚨: Google introduced Gemini World! A universal AI assistant, which will push us forward on the path to AGI. https://t.co/m2y0oMjkmW pic.twitter.com/1uLDpCixV6
— 🚨 AI News | TestingCatalog (@testingcatalog) 20 mai 2025BREAKING : Google introduced Gemini World! A universal AI assistant, which will push us forward on the path to AGI.
-

Google DeepMind Launches Deep Thinking Model at I/O
By
–
It's been a wild run over the last 2 weeks to land Deep Thinking just in time for I/O! Heroic efforts across many teams @GoogleDeepMind from research to production that felt almost impossible last month! Proud of our research team for the relentless effort over the past 6 months
-
Claude 2.5 Pro Token Budgets Coming June 2026
By
–
And thinking budgets coming when 2.5 Pro goes GA in early June, so you can directly control how much the model thinks!
-

4 Design Patterns for Smarter Autonomous AI Agentic Workflows
By
–
Design Patterns for AI Agentic Workflows in LLM Applications
How can AI systems work smarter and more autonomously?
By adopting these 4 design patterns:
• Reflection: Self-assessment and output refinement
• Tool Use: Integrating diverse tools for broader capabilities -

Google introduces Gemini 2.5 Pro Deep Think model outperforms Pro
By
–
BREAKING : Google introduced Gemini 2.5 Pro Deep Think – a new model that outperforms the normal Pro version. It will become available to trusted testers on GCP Vertex via an API.
-

Google Unveils Gemini 2.5 Pro Deep Think Model
By
–
Gemini 2.5 Pro with "Deep Think", our models just keep getting more SOTA, more to share soon : )
-

Google I/O 2025: Major AI Announcements Featuring Gemini Android
By
–
Ready to see Google's next big slate of AI announcements? Google I/O 2025 kicks off today at 1PM ET / 10AM PT. Engadget will be covering #GoogleIO in real-time here and on our live blog: https://
engadget.com/big-tech/googl
e-io-2025-live-updates-on-gemini-android-xr-android-16-updates-and-more-214622870.html
… -
Dynamic NeRF and 4D Representation Optimization Techniques
By
–
The current output is a Dynamic NeRF representation, which doesn’t support explicit formats like FBX or blend shapes. However, one can use the multi-view videos as reference to optimize a different 4D representation.