Google announces Gemini 3 as battle with OpenAI intensifies (www.cnbc.com)

Google announced Gemini 3, its next-generation multimodal AI, positioning it as a response to OpenAI’s rapid advances (GPT-5) and rolling it into the Gemini app, AI Mode/Overviews and enterprise products starting with select subscribers. Google says Gemini 3 needs “less prompting,” better grasps depth, nuance and user intent, and will surface more direct, less sycophantic answers. The model is being deployed where users already are—Gemini app (650M monthly users) and AI Overviews (2B monthly)—and will be available to developers via the Gemini API and Vertex AI for businesses. Technically, Gemini 3 powers “generative interfaces” that combine text, images, tables and interactive elements (examples: custom loan calculators, physics simulations, image-based museum explanations). Google also introduced Google Antigravity, an agent platform that lets developers code at a higher, task-oriented level, and touted Gemini 3 as its strongest “vibe coding” model for code generation. For enterprises, use cases include onboarding, video and factory-image analysis, procurement automation and other domain workflows. The announcement accelerates the competitive arms race with OpenAI, signals continued heavy cloud/infrastructure investment, and suggests a shift toward richer, task-driven multimodal UIs and higher-level agent frameworks that aim to boost developer productivity and reduce prompt engineering.