A new AI model for the agentic era

Note from Google and Alphabet CEO Sundar Pichai:

Information is at the core of human progress. That’s why for more than 26 years we’ve been focused on our mission of organizing the world’s information and making it accessible and useful. That’s why we continue to push the frontiers of artificial intelligence, organizing information from every input and making it accessible through any output so it can actually be useful to you.

this is our vision We launched Gemini 1.0 last December. Gemini 1.0 and 1.5 are the first native multimodal models, making huge strides in multimodality and long-context, allowing them to understand and process information in text, video, images, audio, and code.

Today, millions of developers are developing with Gemini. It helps us reimagine all of our products (including all 7 products with 2 billion users) and create new ones. NotebookLM This is a great example of what multimodality and long context can do for people, and why it’s loved by so many people.

Over the past year, we’ve been investing in developing more agent models, meaning they can learn more about the world around you, think multiple steps ahead, and take action on your behalf under your supervision.

Today, we’re excited to launch the next era of models built specifically for the new agency era: introducing Gemini 2.0, our most powerful model yet. With new advances in multimodality (such as native image and audio output) and the use of native tools, it will allow us to build new artificial intelligence agents, bringing us closer to the vision of a universal assistant.

Today, we’re putting 2.0 into the hands of developers and trusted testers. We’re working quickly to incorporate this into our products, led by Gemini and Search. Starting today, our Gemini 2.0 Flash experimental model will be available to all Gemini users. We’ve also introduced a new feature called In-depth researchwhich uses advanced reasoning and long-context capabilities to act as a research assistant, exploring complex topics and writing reports on your behalf. Available today in Gemini Advanced.

No product is more transformed by artificial intelligence than search. Our AI overview now reaches 1 billion people, empowering them to ask entirely new types of questions – quickly becoming one of our most popular search features ever. Next, we are bringing Gemini 2.0’s advanced reasoning capabilities to AI Overviews to solve more complex topics and multi-step problems, including advanced mathematical equations, multi-modal queries, and encoding. We’re starting limited testing this week and will be rolling it out more broadly early next year. We’ll continue to bring Artificial Intelligence Overview to more countries and languages ​​over the next year.

Advances in 2.0 are the result of a decade of investment in our differentiated, end-to-end approach to AI innovation. It’s built on custom hardware, such as our sixth-generation TPU Trillium. TPU provides 100% support for Gemini 2.0 training and inference, and now Trillium has become generally available to customers so they can build with it too.

If Gemini 1.0 was about organizing and understanding information, then Gemini 2.0 is about making information more useful. I can’t wait to see what the next era brings.

-Sundar