Deep Learning
Introducing Gemma 3: The most powerful model you can run on a single GPU or TPU
To dive deeper into technical details behind these opportunities, as well as a comprehensive overview of our approach to responsible development, can be found...
Deep Learning
Experiment with native image generation in Gemini 2.0 Flash
IN December For the first time, we introduced native image output in Gemini 2.0 Flash to trusted testers. Today we're...
Deep Learning
Gemini Robotics brings artificial intelligence to the physical world
We develop artificial intelligence and robotics responsiblyExploring the continuing potential of artificial intelligence and robotics, we take a multi-layered, holistic approach to...
Deep Learning
Gemini 2.5: Our most wise AI model yet
Last updated March 26Today we introduce Gemini 2.5, our most wise AI model yet. Our first version 2.5 is an experimental version of version...
Deep Learning
Assessing the potential cybersecurity threats of advanced artificial intelligence
Artificial intelligence (AI) has long been a cornerstone of cybersecurity. From malware detection to network traffic analysis, predictive machine learning models and...
Deep Learning
Following a responsible path to AGI
Understanding and addressing potential abuseAbuse occurs when a human intentionally uses an AI system for malicious purposes.Greater insight into contemporary harms and...
Deep Learning
DolphinGemma: How Google’s AI helps decode dolphin communications
Sharing DolphinGemma with the research communityRecognizing the value of collaboration in scientific discovery, we plan to make DolphinGemma available as an open access model...
Deep Learning
Generate Gemini and Whisk videos with Veo 2
Starting today, Gemini Advanced users can generate and share videos using our state-of-the-art video model, I see 2. In Gemini, you can now translate...