Latest publications
SIMA 2 from DeepMind: The First Thinking Agent for Video Games and Robotics
DeepMind presented SIMA 2 — an agent for virtual worlds that evolved from instruction-following to an interactive assistant capable of reasoning about goals, conversing with users, and self-improving.
DeepMind Created AlphaEarth — A Model for Global Earth Monitoring
DeepMind created the AlphaEarth AI model, which combines petabytes of satellite data into a single planetary map with 10-meter resolution and sees through clouds.
Google Created AI Agent AlphaEvolve to Automate Algorithm Design
DeepMind introduced AlphaEvolve — an evolutionary AI agent based on Gemini that automatically designs and optimizes algorithms. Google is already using it to save data center resources, improve chip design, and accelerat
DeepMind Presented Co-Scientist — Multi-Agent AI for Accelerating Scientific Research
DeepMind created a multi-agent AI assistant based on Gemini that helps scientists generate hypotheses and find new solutions for rare diseases, including liver fibrosis treatment.
Google DeepMind Turned the Mouse Cursor Into a Contextual AI Assistant
Google DeepMind introduced a new mouse pointer that understands screen context and helps users right where they work—without switching applications or writing detailed prompts.
Google DeepMind Launches APAC Startup Accelerator Focused on Environmental Innovation
Google launched a three-month accelerator program in the Asia-Pacific region to support startups and research groups developing environmental innovations using artificial intelligence.
Google unveils Gemini 3.5 — a model that automates complex tasks
Google’s Gemini 3.5 Flash executes multi-step workflows automatically. Four times faster than competitors, at half the price. Already used by Shopify, Macquarie Bank, Xero.
WeatherNext Helped National Hurricane Center More Accurately Predict Hurricane Melissa
DeepMind's WeatherNext AI model significantly improved the accuracy of Hurricane Melissa forecasting, giving communities in Jamaica additional time to prepare for the historic natural disaster.
DeepMind Co-Scientist helps geneticists find drugs for cirrhosis
The Co-Scientist tool helped a Stanford geneticist find new ways to treat chronic liver diseases. The research shows how AI can speed up the search for new uses for existing drugs.
DeepMind and MIT Develop RNA Therapy Against Amyotrophic Lateral Sclerosis
Scientists from Boston Children's Hospital and MIT with DeepMind support have developed a new approach to treating amyotrophic lateral sclerosis by combining RNA therapy and computational predictions.
DeepMind's Co-Scientist helps find new ways to treat liver diseases
The AI assistant Co-Scientist helps researchers identify the mechanisms of liver diseases and find new treatments, explaining why some patients respond to drugs differently.
DeepMind Helps Calico Find New Pathways in Aging Research
Google DeepMind has created a Co-Scientist tool that helps Calico Life Sciences find new directions in aging research by connecting fragmented scientific data.
DeepMind Created Co-Scientist to Find Genetic Triggers of New Diseases
DeepMind created Co-Scientist — an AI system for identifying genetic causes of infectious diseases. The tool analyzes data and helps scientists develop vaccines and antiviral drugs faster during outbreaks.
DeepMind introduces tools for tracking content creation and editing
DeepMind is expanding tools to track the provenance of web content and its edits — a new initiative to improve transparency and fight disinformation online.
DeepMind Unveiled Scientific Tools for Gemini: Hypothesis Generation and Experiment Automation
DeepMind announced three Gemini-based tools for scientific research: a hypothesis generator, a system for parallel code testing, and scientific literature analysis. The tools reduce analysis time from hours to minutes.
Google Introduces Gemini Omni Flash — Model for Creating Video from Text and Images
Google launched Gemini Omni Flash — an AI model that creates high-quality video content from text, images, and audio recordings, edits video through natural language, and is already available in Gemini app and YouTube Sh
DeepMind's Project Genie Learns to Simulate Real Places from Google Street View
DeepMind has expanded Project Genie access to Google AI Ultra subscribers worldwide and integrated the ability to simulate real places through Street View.
How DeepMind's Co-Scientist Helps Rejuvenate Human Cells
Biologists used DeepMind's Co-Scientist AI system to discover new genetic factors that successfully rejuvenate human cells and may pave the way for treating age-related diseases.
Google DeepMind and Singapore Begin Collaboration in Frontier AI
Google DeepMind has partnered with Singapore to develop frontier AI solutions in healthcare, education, and sustainable development.
DeepMind unveils AlphaEvolve: a Gemini-powered AI agent for development and science
DeepMind has launched AlphaEvolve, a Gemini-based AI agent that automates development and is used in business, infrastructure, and scientific research, scaling results across different sectors.
Google DeepMind proposed a cognitive framework for measuring progress toward AGI
Google DeepMind released a scientific framework for assessing progress toward AGI and launched a Kaggle hackathon inviting researchers to build missing cognitive tests for AI.
Google DeepMind Introduces AI co-clinician System for Doctors and Telemedicine
Google DeepMind launched a research AI co-clinician system that helps doctors with clinical questions and is being tested in telemedicine, but is not intended for diagnosis or treatment.
Google DeepMind Introduces Gemini Robotics-ER 1.6 for Autonomous Real-World Tasks
Google DeepMind released Gemini Robotics-ER 1.6 — a robot model with improved spatial reasoning, multi-camera analysis, and industrial gauge reading capabilities.
Google DeepMind and South Korea Launch Partnership for Scientific Breakthroughs with AI
Google DeepMind, together with South Korea's Ministry of Science, will open an AI Campus in Seoul and provide scientists access to models for biology, climate, energy, and new research.
Gemini 3.1 Pro: Google bets on complex tasks
Google unveiled Gemini 3.1 Pro — a model built for tasks where a simple answer is not enough. A new level of intelligence for the most demanding scenarios.
Google DeepMind launches AI initiative for science and education in India
DeepMind is expanding its national partnerships program: India will become a new hub for AI research and educational technology.
Gemini can now create music from text and photos
Google integrated the Lyria 3 model into Gemini — now any user can generate 30-second tracks from a description or an image.
Project Genie: Google Hands Over the Keys to Infinite Game Worlds
While everyone is debating how realistically Sora draws cat fur, Google decided to come out swinging and give users what we've been so desperately missing in generative video—control.