Google has announced the launch of Gemini 3, described as its most intelligent and capable Artificial Intelligence (AI) model, designed to deliver state-of-the-art reasoning and powerful new agentic capabilities across Google products used daily around the world.
The Chief Executive Officer of Google and Alphabet, Sundar Pichai, disclosed this in a statement on Tuesday, noting that the launch marks a major milestone in Google’s mission to make AI more helpful and accessible to everyone.
Pichai said the new chapter in AI innovation would see Google continue to advance intelligent agents and personalized experiences.
According to him, the rollout of Gemini 3 comes at a time when global interest in AI tools is rising. He revealed that in Nigeria, Google Search interest in “AI and studying” has grown by over 200 per cent, while queries for “how to learn AI” have increased by 160 per cent in the past year.
“A new era of intelligence begins with the Gemini 3 model itself, which is built to grasp depth and nuance like never before,” he said.
Pichai explained that Gemini 3 Pro, the first model in the new family, demonstrates PhD-level reasoning and achieves new performance records on major AI benchmarks. These include complex problem-solving—scoring 91.9 per cent on the GPQA Diamond benchmark—and advanced understanding of videos, images, and handwritten text.
He added that the leap in core intelligence underpins the new experiences introduced with the model, bringing smarter, more direct responses that favour genuine insight over generic output.
Pichai said Gemini 3 also excels in “vibe coding,” enabling developers to bring ideas to life using natural language, boosted by enhanced agentic capabilities that allow the system to better understand intent, plan multi-step tasks, and act with minimal user prompting.
He said these advancements power more interactive experiences within Google Search, transforming it from a traditional information tool into a dynamic learning companion. For the first time, a new Gemini model will be available in Search from day one in AI Mode.
According to him, Gemini 3 can generate dynamic visual layouts, interactive tools, and real-time simulations. “Instead of simply reading about physics, users can now interact with a simulation of the three-body problem,” he said. Similarly, users exploring mortgage options could access a custom-built loan calculator generated instantly for side-by-side comparisons.
The CEO said the evolution continues with a new agent in the Gemini app, which becomes a personal assistant capable of taking action on behalf of users. For Google AI Ultra subscribers, the experimental Gemini Agent can coordinate and complete complex multi-step tasks across connected Google apps.
Pichai also announced the launch of Google Antigravity, a new agentic development platform designed to help creators build advanced AI-driven experiences. The platform transforms AI from a simple coding assistant into an active collaborator that can autonomously plan and execute complex software tasks—from building features to fixing bugs.
“Google Antigravity is built on core tenets of trust, autonomy, and feedback,” Pichai said. “It functions as a mission control for orchestrating multiple agents, giving developers task-level context and verifiable artefacts to ensure confidence in the work being done.”
He added that this shift to an agent-first approach represents a fundamental change in how software will be developed, enabling creators to operate at a more strategic and innovative level.
