Google Launches Gemini 3, Ushering in a New Era of Agentic AI for Kenya
A Major Leap in Intelligence and Capability
Google has unveiled Gemini 3, its most advanced AI model to date, bringing state-of-the-art reasoning and powerful new agentic abilities to millions of users.
The launch marks a major shift in how Google products will work, with AI becoming more intuitive, interactive, and helpful for people in Kenya and across Africa.
Sundar Pichai, CEO of Google and Alphabet, said the company is entering a new chapter focused on intelligence, agents, and personalisation.
Kenya stands out as a prime adopter, with Google Search data showing a 150% surge in interest in AI over the past year, and searches for “AI jobs” more than tripling. The appetite for advanced tools continues to rise as Kenyans look to tech to power careers and new economic opportunities.
Gemini 3 Pro, the first release in the new family, is designed to handle reasoning with greater depth and nuance. It delivers PhD-level problem-solving and sets new records on major AI benchmarks, including a 91.9% score on GPQA Diamond.
The model can understand video, images, diagrams, and even handwritten notes, enabling more accurate and meaningful responses.
Smarter Search, Stronger Tools, and New Agentic Power
Gemini 3 Pro also brings a more natural interaction style, offering clearer and more insightful responses. It supports “vibe coding,” allowing developers to turn ideas into working prototypes using simple language. Its new agentic abilities help it understand intent, plan tasks, and complete multi-step actions with minimal user input.
These upgrades enable a new experience in Google Search. For the first time, a Gemini model is available in Search on day one. In AI Mode, the model creates dynamic visual layouts, simulations, and tools based on the user’s query.
A learner can interact with a physics simulation, while someone researching a mortgage can receive a customised loan calculator instantly.
Google is also upgrading the Gemini app with a more capable personal agent. For users subscribed to Google AI Ultra, the new experimental Gemini Agent can perform complex tasks across Google apps. It can organise an inbox, prioritise tasks, draft responses, compare rental options, or prepare bookings all while leaving the user in full control.
A New Platform for Developers: Google Antigravity
To help builders create their own intelligent agents, Google has introduced Antigravity, an agentic development platform. It acts as a mission control system where developers can orchestrate multiple agents to plan, build, and fix software automatically.
The platform is grounded in trust, autonomy, and feedback. It gives developers clear oversight, context, and verifiable outputs, allowing AI to handle the heavy lifting while humans steer the vision. This shift moves software creation from manual coding to strategic supervision.
Google says Gemini 3 has passed its most extensive safety evaluations yet. Gemini 3 Pro is available starting today through the Gemini app and to Google AI Pro and Ultra subscribers within AI Mode in Search.
