Over the last decade, we have laid many of the foundations for the modern era of artificial intelligence, starting with pioneering solutions Transformer the architecture on which all major language models are based, to the development of agent systems that can learn and plan, such as AlphaGo and AlphaZero.
We have applied these techniques to make breakthroughs in quantum computingmathematics, life sciences and algorithmic discoveries. We continue to double the breadth and depth of our fundamental research as we work to discover the next big breakthroughs needed for artificial general intelligence (AGI).
That's why we're working to expand our best multimodal base model, Gemini 2.5 Pro, to become a “world model” that can plan and imagine new experiences by understanding and simulating aspects of the world just like the brain does.
We've been moving in this direction for some time, from our pioneering job training agents, to mastering complex games like Go and StarCraft, to building Genie 2, which can generate simulated 3D environments that you can interact with from a single image prompt.
We are already seeing evidence of these abilities emerging in Gemini's ability to use worldly knowledge and reasoning to represent and simulate the natural environmentVeo's deep understanding of intuitive physics and how Gemini Robotics teaches robots to grasp, follow instructions and adapt on the fly.
Making Gemini a model of the world is a key step in developing a new, more general and useful type of artificial intelligence – a universal AI assistant. It's intelligent AI that understands the context you're in and can plan and take actions on your behalf on any device.
