Gemini 3 brings your ideas to life
Gemini is one of the most widely used AI models right now thanks to its coupling in the Android ecosystem and Google Search. What users can do with the new Gemini 3 model is have an interactive web page generated from a piece of content. With this, instead of having to read a long piece of text, the information is presented as a fully interactive webpage. Google itself calls this technique Dynamic View.
Another interesting addition is "Visual Layout" where Gemini 3 can use images and tables to give a magazine-like response. That way, you can instantly see what the AI model has come up with when you ask it to plan a day trip to, say, Rome.
Furthermore, Google is also working on an experimental Agent that you can put to work cleaning up your mailbox, for example. But it can also prepare draft messages for still unanswered emails. For now, the Agent can only be used in the US with an AI Ultra subscription.
Gemini 3 Pro can bring any idea to life thanks to improvements in reasoning. In a number of benchmarks, Google's new model shows that it can hold its own against other well-known AI models Claude Sonnet 4.5 from Anthropic and GPT5.1 from OpenAI. In the Gemini app, you can now select Gemini 3 as a model.
