Overview
- Gemini 3 Pro is live today in the Gemini app for all users and in Google Search’s AI Mode for AI Pro and Ultra subscribers in the U.S., accessible via the “Thinking” model picker.
- The release introduces generative UI experiments that can build dynamic layouts and interactive tools, with Search responses now able to generate visuals, tables, grids, and real-time simulations.
- Google is rolling out agentic capabilities, including an experimental Gemini Agent for multi‑step tasks in the Gemini app and Antigravity, a new IDE where AI agents operate across editor, terminal, and browser.
- Developers and enterprises can use Gemini 3 via AI Studio, Vertex AI, the Gemini API/CLI, and select third‑party integrations such as Cursor, GitHub, JetBrains, Manus, and Replit.
- Google touts benchmark gains with Gemini 3 Pro topping LMArena and leading WebDev Arena, as a higher‑reasoning Deep Think mode enters safety testing ahead of availability for AI Ultra subscribers in the coming weeks.