Google has introduced the Gemini 3 AI Modelan update to Gemini with improved visual reasoning, and Google Antigravityan agent development platform for AI-assisted software development.
Both were announced on November 18. Gemini 3 is positioned to deliver reasoning capabilities with robust function calls and instruction fulfillment to create sophisticated agents. Agent capabilities in Gemini 3 Pro are integrated into agent experiences in tools like Google AI Studio, Gemini CLI, Android Studio, and third-party tools. Multimodal reasoning and generation allow developers to go from concept to a functional application, making Gemini 3 Pro suitable for developers of any experience level, Google said. Gemini 3 outperforms Gemini 2.5 Pro in coding, mastering both agent workflows and complex scratch tasks, according to the company. Gemini 3 is available in preview for $2 per million tokens in and $12 per million tokens out for requests of 200,000 tokens or less through the Gemini API in Google AI Studio and Vertex AI for enterprise.
Meanwhile, Google Antigravity is moving from agents built into tools to an AI agent as the primary interface. Manage surfaces such as the browser, editor, and terminal to complete development tasks. While the core is a familiar AI-powered IDE experience, Antigravity is evolving the IDE toward an agent-first future, with browser control capabilities, asynchronous interaction patterns, and an agent-first product form factor, according to Google. Together, these enable agents to autonomously plan and execute complex end-to-end software tasks. Available in public preview, Antigravity is compatible with Windows, Linux, and macOS.
#Google #unveils #Gemini #model #antigravity #agent #development #tool
