Google enters ‘new era’ of AI with its advanced Gemini 3 model
Google launches Gemini 3, its smartest AI model yet. Right off the bat, users will have access to the flagship Gemini 3 Pro model, which is multimodal out of the box and can process text, images, audio and video in the same stream. This model is claimed to outperform several AI benchmarks in logic, mathematics, and fact-checking. It is also said to provide shorter, more direct and less biased answers.
A new feature is that Gemini 3 can generate interfaces as presentations or web layouts based on user instructions. There is also an improved search functionality with expanded “query streaming” that will allow the model to better understand the intent behind a question.
In the Gemini app and Google Search’s AI mode, answers are now also displayed in a way that combines visual elements such as images, tables, grids and even interactive simulations.
On top of that, Gemini 3 Pro has gained sharper reasoning and planning capabilities, allowing it to handle more complex and longer-term tasks. It powers Google’s new experimental tool, Gemini Agent, which can, for example, organize emails or book trips for users.
Gemini 3 Pro is now available in the Gemini app in preview form. In the United States, Google AI Pro and Ultra subscribers can also test the Agent feature as well as the new Advanced AI mode.
