Google's Gemini 3 and Antigravity: The new era of AI intelligence
Unlike other AI chatbots that only exist on screen and ask questions, Antigravity can understand existing code, interpret what developers are trying to build, and generate the code for them
Just imagine a world where a non-developer can build a wonderful app that serves millions of people.
On 18 November, Google CEO Sundar Pichai unveiled the new version of Google's generative AI, Gemini 3. They call it the most advanced AI model to date.
With Gemini 3, anyone can learn even complicated topics using advanced visualisation and comprehensive learning techniques. Google claims it is the "New Era of Intelligence."
Keep updated, follow The Business Standard's Google news channel
Gemini 3 is rolling out for general users, software professionals and developers even entrepreneurs. General users can access from the Gemini App, and for Google AI Pro and Ultra subscribers in AI Mode in Search, entrepreneurs can use it via Vertex AI and Gemini Enterprise (Vertex AI is Google Cloud's unified and fully-managed platform for building, deploying, and managing machine learning models).
Demis Hassabis, CEO of Google DeepMind, and Koray Kavukcuoglu, CTO of Google DeepMind and chief AI architect at Google, are the visionaries consistently improving Gemini every year.
Google also released a new coding platform called Antigravity, the most powerful vibe coding agent right now. Gemini 3 includes agentic features, meaning anyone can build complicated software and applications from just a prompt.
However, Antigravity marks a new era for non-developers. Unlike other AI chatbots that only exist on screen and ask questions, Antigravity has groundbreaking features.
It can understand existing code, interpret what developers are trying to build, and generate the code for them. Antigravity can also test the code and detect problems without human intervention. It is the new era for developing own applications/tools.
Gemini 3 is the most powerful reasoning AI tool yet and excels at spatial reasoning (understanding layouts, trajectories), which is especially useful in robotics and augmented reality (AR). It supports 1,000,000 tokens of content and can deliver outputs of up to 640,000 tokens.
The new search integration sets another benchmark for the Google ecosystem. While Gemini 2.5 could build heavily optimised generative images and videos, Gemini 3 can produce even more feasible generative images.
Gemini 3 is also extremely powerful for reasoning, analysis, and analytics. According to Google, Gemini 3 Pro maintains consistent tool usage and decision-making for a full simulated year of operation, delivering higher returns without drifting off-task.
Google has also introduced Gemini 3 Deep Think, with enhanced reasoning capabilities and improved performance. Deep Think achieved 41% on Humanity's Last Exam without tools. It can analyse any data quickly and accurately, making it a powerful assistant for data analysts.
Deep Think will only be available to Google AI Ultra subscribers. Although Gemini offers a free model, the most advanced features are available only in the paid models: Google AI Pro ($20 per month) and Google AI Ultra ($250 per month).
