Deep Learning
Genie 3: A fresh border for global models
ThanksGenie 3 was possible because of the key cartridges of research and engineering Phila Ball, Jakub Bauer, Frank...
Deep Learning
Thinking about the way we measure the intelligence of AI
Current AI comparative tests are trying to keep up with state-of-the-art models. Although they are helpful in measuring the performance of the model in...
Deep Learning
Try a deep thought in the Gemini application
How deep thinking works: extension of the parallel "thinking time" GeminiLike people deal with sophisticated problems, devoting time to examining different sides, considering potential...
Deep Learning
Gemini 2.5 Flash-Lite is now stable and generally available
Today we release the stable version of Gemini 2.5 Flash-Lite, our fastest and lowest cost (0.10 USD per 1 m,...
Deep Learning
The advanced version of Gemini from Deep Think officially reaches the gold standard at the International Mathematical Olympics
ThanksWe thank the International Organization of the Mathematical Olympiad for their support.Thang Luong managed the general technical direction...
Deep Learning
We are expanding our family of Gemini 2.5 models
We designed Gemini 2.5 to be a family of hybrid reasoning models that ensure incredible performance, and at the same time be in Pareto...