Google is turning up the heat in the AI race. And this time, it’s shipping its newest brain straight into Search from day one.
The tech giant on Tuesday launched Gemini 3, a major upgrade to its AI lineup that’s rolling directly into its biggest products, including the Gemini app and Google Search. The company says the AI is its most powerful model yet, delivering deeper reasoning, sharper understanding, and new interactive capabilities.
Google hopes these new features will give it an edge in the accelerating race with OpenAI, Anthropic, and others.
Alphabet CEO Sundar Pichai described the moment as a turning point for the company’s AI roadmap. In a blog post, he called Gemini 3 “our most intelligent model” and added that the system is “built to grasp depth and nuance.” Pichai noted that it is “much better at figuring out the context and intent behind your request, so you get what you need with less prompting.”
Pichai also highlighted how far Google has come since the first Gemini release nearly two years ago, saying Gemini is “one of our biggest scientific and product endeavors ever undertaken as a company.”
He noted that “AI Overviews now have 2 billion users every month. The Gemini app surpasses 650 million users per month, more than 70% of our Cloud customers use our AI, 13 million developers have built with our generative models, and that is just a snippet of the impact we’re seeing.”
What Gemini 3 can do
Google says Gemini 3’s biggest leap is in reasoning.
Demis Hassabis, CEO of Google DeepMind, emphasized that the model now provides answers with less empty praise, saying it delivers responses that are “trading cliché and flattery for genuine insight — telling you what you need to hear, not just what you want to hear.”
Benchmarks show big jumps, too. Google reported that “it tops the LMArena Leaderboard with a breakthrough score of 1501 Elo. It demonstrates PhD-level reasoning with top scores on Humanity’s Last Exam (37.5% without the usage of any tools) and GPQA Diamond (91.9%).”
“It also sets a new standard for frontier models in mathematics, achieving a new state-of-the-art of 23.4% on MathArena Apex,” Google added.
Gemini 3 in search: A visual, interactive upgrade
Search is getting a makeover powered by Gemini 3. The advanced reasoning power of the new model enables Search to understand complex queries better and find more relevant content.
Beyond text, it unlocks new Generative UI experiences. This means Search can dynamically create custom visual layouts, interactive tools, and simulations tailored to your question.
For instance, if you’re trying to grasp the complex physics of the three-body problem, the model could generate a real-time, interactive simulation right in your search results. Elizabeth Hamon Reid, VP of Engineering for Search, explained this shift: “When the model detects that an interactive tool will help you better understand the topic, it uses its generative capabilities to code a custom simulation or tool in real-time and adds it into your response.”
A new coding era with Google Antigravity
Alongside the main model, Google launched Google Antigravity, a new agent-powered coding environment that blends a chat window, terminal, and browser.
“It’s a faster way to develop: you act as the architect, collaborating with intelligent agents that operate autonomously across the editor, terminal, and browser,” said Logan Kilpatrick, Product Lead, Google AI Studio and the Gemini API.
The platform enables AI agents to write, run, and validate code as they work, an early look at Google’s vision for more autonomous development tools.
“These agents plan and execute complex software tasks, communicating their work with the user via detailed artifacts. This elevates all aspects of development, from building features, UI iteration, and fixing bugs to researching and generating reports,” he noted.
Gemini agent for everyday tasks
Beyond coding, Google is also launching Gemini Agent, an experimental assistant that performs multi-step tasks across apps like Gmail and Calendar. Josh Woodward, VP of Google Labs, said the system lets users offload work such as inbox cleanup or travel planning.
In his blog, Woodward explained that Gemini Agent “handles multi-step tasks directly inside Gemini” and can do things like “organize my inbox” or compare travel options using details from email. The tool will be available first to Google AI Ultra subscribers in the US.
Enterprises also get Gemini 3 through Vertex AI, while developers can tap into the model via the Gemini API in AI Studio. The preview price sits at $2 per million input tokens and $12 per million output tokens, according to Google.
Google says the model is already powering features across its ecosystem, from multimodal analysis to document understanding and long-form video reasoning.
Google races to keep up and stand out
The launch arrives as Google faces fierce competition from OpenAI, which recently pushed out GPT-5.1, and Anthropic’s latest Claude 4.5 Sonnet.
But with Gemini 3 instantly powering revenue-generating products like Search, Gmail integrations, and enterprise tools, Google is betting that speed and scale are its biggest advantages.
“We think Gemini has set quite a new pace in terms of both releasing the models, but also getting it to people faster than ever before,” said Google’s chief AI architect, Koray Kavukcuoglu, per Reuters.
For a deeper look at Google’s latest breakthroughs, don’t miss our coverage of WeatherNext 2 and how it’s redefining AI-powered forecasting. Read it here.