Google DeepMindHow a Gemma model helped discover a new potential cancer therapy pathway
Exploring the role of Gemma models in uncovering new therapeutic pathways for cancer treatment.
Google DeepMindIntroducing Veo 3.1 and advanced creative capabilities
Discover the new features and advanced creative tools in Veo 3.1 that enhance user experience and productivity.
Google DeepMindIntroducing the Gemini 2.5 Computer Use model
An introduction to the Gemini 2.5 Computer Use model, outlining its key concepts and applications.
Google DeepMindVaultGemma: The world's most capable differentially private LLM
Introducing VaultGemma, an advanced large language model with cutting-edge differential privacy features ensuring secure and private data handling.
Google DeepMindImage editing in Gemini just got a major upgrade
Gemini introduces significant advancements in image editing capabilities, enhancing user experience and functionality.
Google DeepMindIntroducing Gemma 3 270M: The compact model for hyper-efficient AI
Discover the capabilities and efficiency of the compact Gemma 3 270M AI model designed for advanced machine learning tasks.
Google DeepMindRethinking how we measure AI intelligence
Exploring innovative approaches to evaluating AI intelligence beyond traditional metrics.
Google DeepMindTry Deep Think in the Gemini app
Explore the innovative Deep Think feature integrated into the Gemini app to enhance user experience.
Google DeepMindBringing AI to the next generation of fusion energy
Leveraging AI and advanced simulations like TORAX, the partnership between Google DeepMind and Commonwealth Fusion Systems accelerates the optimization and control of fusion energy in the SPARC tokamak to achieve sustainable, net-positive fusion power.
Google DeepMindIntroducing CodeMender: an AI agent for code security
CodeMender is an AI-powered autonomous agent leveraging advanced deep learning models to detect, patch, and proactively secure software vulnerabilities through automated reasoning, debugging, and validation techniques.
Google DeepMindGemini Robotics 1.5 brings AI agents into the physical world
Gemini Robotics 1.5 introduces advanced AI models enabling robots to perceive, plan, reason, and act autonomously in complex physical environments with enhanced safety and generalization across embodiments.
Google DeepMindStrengthening our Frontier Safety Framework
The updated Frontier Safety Framework enhances risk assessment and mitigation for advanced AI by introducing Critical Capability Levels to address harmful manipulation, misalignment, and destabilizing AI research, ensuring responsible development toward AGI.