Google announces Gemini 3 as battle with OpenAI intensifies

Google has officially introduced its latest artificial intelligence model, Gemini 3, intensifying its competition with OpenAI, the creator of ChatGPT. This advanced AI is designed to provide users with improved answers to complex inquiries, requiring less prompting to deliver the necessary information, as noted by Alphabet CEO Sundar Pichai in a series of blog posts released on Tuesday. Gemini 3 will be seamlessly integrated into various platforms, including the Gemini app, AI Mode, and AI Overviews, as well as Google's enterprise solutions. The rollout commenced on Tuesday for select subscribers, with a broader launch expected in the weeks ahead. This announcement follows the launches of previous models, Gemini 2.5 and Gemini 2.0, which debuted eight and eleven months ago, respectively. Pichai highlighted the rapid evolution of AI, stating, "It's amazing to think that in just two years, AI has evolved from simply reading text and images to reading the room." He emphasized that Gemini is now operational at the scale of Google, boasting 650 million monthly active users for the app and 2 billion for AI Overviews, while OpenAI reported 700 million weekly users for ChatGPT as of August. The new model aims to excel in understanding depth and context, with Pichai asserting that it is much better at discerning the intent behind user requests. Google plans to continue utilizing its other AI models for simpler tasks. In a broader context, major tech companies, including Alphabet, Meta, Microsoft, and Amazon, are heavily investing in AI infrastructure, collectively forecasting capital expenditures to exceed $380 billion this year. According to Demis Hassabis, CEO of Google's AI unit DeepMind, responses generated by Gemini 3 will prioritize genuine insights over flattery, aiming to provide users with the information they need to hear rather than what they want to hear. Industry experts have critiqued existing AI chatbots for being overly agreeable. In addition to Gemini 3, Google introduced a new agent platform named "Google Antigravity," which allows developers to code at a higher, task-oriented level. This latest AI model is described as the company's "best vibe coding model ever," facilitating the creation of code through prompts. Google's vision for Gemini 3 includes the development of "generative interfaces," which will present answers in a visually engaging format akin to a digital magazine. For instance, when tasked with explaining the Van Gogh Gallery, Gemini provided a vibrant, image-based overview of each painting. Initially, Gemini 3 will be available to paid subscribers within AI Mode, where it will analyze questions and generate layouts featuring visual elements like images and tables. The model promises to create interactive tools, such as a custom loan calculator or simulations for complex physics problems. Developers can access the Gemini API, enabling businesses to integrate it via Vertex AI, Google's cloud service for managing AI models. For corporate clients, Gemini 3 offers capabilities such as improving employee training, accurately analyzing videos and images from factory floors, and streamlining procurement processes.