Google Launches Gemini 3: The New "Smartest" Artificial Intelligence Model
Google officially launches Gemini 3 – a new series of artificial intelligence models. As noted by the company, these systems are currently the "smartest" and "most accurate".
What is Known
Gemini 3 Pro is a multimodal model capable of processing text, images, and audio simultaneously. The company provides use cases: converting photo-based recipes into an electronic cookbook, creating interactive flashcards from video lectures. The Gemini app features tools for creating complete programs and interfaces, including the Canvas tool and a new "generative interfaces" feature.
In AI mode in Google Search, the model generates visual responses, including tables, images, grids, and simulations. The updated query distribution system better handles complex questions and clarifies user intent.
Gemini 3 Pro received changes in response style: Google notes that responses have become shorter, more informative, and content-oriented. The model also showcases updated reasoning and planning capabilities, as well as agent functions – it can perform tasks in an app, such as analyzing emails or preparing travel information.
The model has already appeared in the Gemini app. Users in the US with a subscription to Google AI Pro and AI Ultra can activate Gemini 3 Pro in AI Search mode. The Gemini Agent feature is currently available to AI Ultra subscribers.
Source: Google
