Google has unveiled Gemini 3, the company’s most advanced artificial intelligence model to date, marking a pivotal moment in the evolution of AI capability, scale and everyday usefulness. Announced by Sundar Pichai, Chief Executive Officer of Google, the new model represents a profound step forward in reasoning, contextual awareness and multimodal intelligence.

Gemini 3 brings together the full suite of Gemini abilities into a single cohesive model designed to help users bring any idea to life with clarity and depth. From decoding complex problems to understanding the subtleties of creative expression, Gemini 3 is engineered to grasp intent, nuance and context like never before.

“AI has grown remarkably in just two years, moving from reading text and images to something far more human: reading the room,” said Sundar Pichai. “Gemini 3 is our most intelligent model yet. It is built to understand deeper intent, interpret subtle signals and help people achieve better results with less prompting. We cannot wait to see what people build with it.”

Built for sophisticated reasoning and natural understanding

Gemini 3 introduces state of the art reasoning capabilities that allow it to unpack layered challenges, interpret intricate patterns and support users through tasks that demand both precision and imagination. Its enhanced contextual understanding dramatically reduces the need for detailed instructions, creating a more natural and intuitive experience for users, developers and enterprises.

Rolling out across Google products from day one

True to the ambition of delivering AI at the scale of Google, Gemini 3 is being deployed across the company’s ecosystem from the moment of launch. This includes:

• AI Mode in Search, offering more advanced reasoning and fluid dynamic experiences

• The Gemini app, enabling users to access the new model directly

• AI Studio, giving developers a more powerful foundation for experimentation

• Google Antigravity, the company’s new agentic development platform

• Vertex AI and Gemini Enterprise, bringing enhanced capability to businesses globally

This marks the first time Google has launched a new Gemini model in Search on the very first day, underscoring the company’s commitment to rapid deployment and practical utility.

A milestone shaped by global teams

Pichai expressed deep gratitude to the engineering, research and product teams across Google who delivered the model, noting that Gemini 3 reflects years of focused innovation and close collaboration across disciplines.

“This is a big day for everyone who has helped bring Gemini 3 to life,” he added. “We will continue to improve it and expand how it can help people and organisations do more. There is much more to come.”

As Google continues to refine the model and extend its capabilities, Gemini 3 is expected to accelerate the pace of AI creativity, productivity and problem solving worldwide.