Google has introduced Gemini 3, an artificial intelligence (AI) tool the company describes as its most advanced model so far.
The launch was announced on Tuesday, with Google highlighting the model as a major step forward in intelligence and reasoning, marking the beginning of a new era of agentic AI across its products.
Sundar Pichai, chief executive officer (CEO) of Google and Alphabet, said that in this “new chapter, we’ll continue to push the frontiers of intelligence, agents, and personalisation”.
Google noted that interest in AI continues to rise globally, pointing to Nigeria’s more than 200 percent increase in Google searches for “AI and studying,” and a 160 percent increase in queries about “how to learn AI” in the past year.
The company said the Gemini 3 family debuts with Gemini 3 Pro, which demonstrates stronger comprehension and advanced reasoning skills.
According to Google, the model sets new performance records on various industry benchmarks, scoring 91.9 percent on the GPQA Diamond test for complex problem-solving and performing exceptionally well in interpreting images, videos, and handwritten text.
Google explained that Gemini 3 Pro offers more accurate and insightful results, designed to act as a collaborative partner rather than a typical assistant.
The company also highlighted the model’s improved ability to carry out multi-step tasks, better understand user intent, and complete tasks with fewer prompts.
In addition, Google said developers can more easily turn ideas into functional code using the model, noting that it excels at “vibe coding”.
Google stated that the technology is now integrated into Google Search, creating a more interactive and learning-oriented experience.
The company added that Gemini 3 can automatically produce dynamic visual layouts, simulations, and tools customized to a user’s query.
For instance, users learning physics can engage with a three-body problem simulation, while those researching mortgages can access a custom loan calculator generated in real time.
Google also launched an upgraded agent in the Gemini app capable of executing tasks across Google’s ecosystem.
For Google AI Ultra subscribers, the experimental Gemini Agent can organize emails, draft replies, and complete multi-step tasks such as finding travel details or comparing rental car options.
A prompt like “organise my inbox” allows the agent to prioritize messages and prepare responses, while “help me book a mid-size SUV for my trip” enables it to gather flight information, compare car rentals, and complete the booking process.
To support developers, Google introduced Google Antigravity, a new agentic development platform designed to autonomously plan and execute software tasks.
The company described it as a “mission control” system for coordinating multiple AI agents, ensuring transparency, verifiable outputs, and developer oversight.
Google said this marks a shift toward an agent-first approach to software development, allowing creators to focus on strategy while AI handles complex operations.
The company added that Gemini 3 underwent its most extensive safety reviews to date.
Google confirmed that Gemini 3 Pro is now available in the Gemini app and also accessible to Google AI Pro and Ultra subscribers through AI Mode in Search.