Google has officially launched Gemini 3, its most advanced artificial intelligence model to date, signalling what CEO Sundar Pichai calls “another big step on the path toward AGI.” The announcement comes nearly two years after Google introduced the Gemini era — an ambitious scientific and product journey that has rapidly reshaped the company’s global AI footprint.
A Breakneck Rise: Billions Now Using Gemini-Powered Products
In a note shared by Pichai, Google highlighted the explosive growth and adoption of its AI ecosystem:
-
AI Overviews: now reach 2 billion users every month
-
Gemini App: more than 650 million monthly users
-
Google Cloud: over 70% of customers use Google’s AI tools
-
Developers: 13 million have built with Google’s generative models
Pichai emphasized that this momentum is powered by Google’s “differentiated full-stack approach,” spanning world-leading infrastructure, research, models, and global-scale products.
From Gemini 1 to Gemini 3: Rapid Evolution in Two Years
Pichai noted how each generation of Gemini has introduced breakthrough capabilities:
-
Gemini 1 unlocked native multimodality and unprecedented long context windows.
-
Gemini 2 laid the groundwork for agentic behavior and advanced reasoning — with Gemini 2.5 Pro dominating the LMArena benchmark for over six months.
-
Gemini 3, launched today, brings all prior innovations together into what Google is calling its most intelligent, context-aware, and nuanced AI model yet.
“It’s amazing to think that in just two years, AI has evolved from simply reading text and images to reading the room,” Pichai said.
Gemini 3: Deeper Reasoning, Better Context, and More Human-Like Interactions
According to Google DeepMind CEO Demis Hassabis and CTO Koray Kavukcuoglu, Gemini 3 sets a new global standard:
-
Superior multimodal understanding across text, images, audio, and video
-
Powerful agentic capabilities for planning, problem-solving, and dynamic interaction
-
Advanced “vibe coding” for natural, expressive, conversational responses
-
State-of-the-art reasoning performance, now enhanced further through Gemini 3 Deep Think
The new model is designed to grasp deeper intent, navigate layered problems, and generate richer visualizations — all while requiring far less prompting from users.
Rolling Out Across Google Products — On Day One
For the first time ever, Google is deploying its newest model across Search immediately.
Gemini 3 is now available in:
-
AI Mode in Google Search, enabling complex reasoning and dynamic results
-
The Gemini app
-
AI Studio and Vertex AI for developers
-
Google Antigravity, a new agentic development platform
Google is also previewing Gemini 3 Pro, giving users early access to the model’s capabilities, and preparing Gemini 3 Deep Think for safety testers before expanding availability to AI Ultra subscribers.
Pushing Toward AGI — With Safety at the Core
Hassabis and Kavukcuoglu describe Gemini 3 as the start of a new era focused on intelligence, agents, and personalization. While the company remains committed to pushing boundaries, the rollout will continue to be phased with safety testing and oversight.
A New Chapter in AI
As Pichai closed his note:
“We hope you like Gemini 3… and look forward to seeing what you build with it. Much more to come!”
With Gemini 3, Google is not just improving AI — it’s positioning itself at the center of a rapidly accelerating race toward general intelligence, reshaping how billions of people learn, build, work, and create.
