Overview
- Gemini 3 Pro is available to all users in the Gemini app, bringing richer visual and structured responses across features like Canvas and Gemini Labs.
- Google says the model natively processes text, images and audio in a single pass, enabling outputs such as recipe books from photos and interactive study cards from videos.
- The company highlights stronger reasoning and agent capabilities with stable long‑term planning, while a Deep Think mode remains limited to safety testers.
- Gemini 3 Pro currently sits at the top of the LMArena benchmark, underscoring Google's claimed advances in performance.
- Developers can use the Pro preview in AI Studio and find it in Vertex AI as "gemini-3-pro-preview-11-2025", with Gemini agents initially rolling out to AI Ultra subscribers.