Google is today starting to launch Gemini 3, a new series of models the company says are its “most intelligent” and “factually accurate” AI systems to date. They’re also a chance for Google to take a leap ahead of OpenAI following the stellar launch of GPT-5, which will potentially put the company at the forefront of consumer-focused AI models.
For the first time, Google is giving everyone access to its new flagship AI model – Gemini 3 Pro – in the Gemini app on day one. It is also offering Gemini 3 Pro for customers inside Search. Tulsi Doshi, senior director and product lead at Google DeepMind, says the new model will bring the company closer to making information “universally accessible and useful” as its search engine continues to evolve.
“I think a really big step in that direction is to break out of the paradigm of just text responses and give you a richer, more complete view of what you can actually see.”
The Gemini 3 Pro is “natively multimodal”, meaning it can process text, images and audio all at once, rather than handling them separately. For example, Google says the Gemini 3 Pro can be used to translate photos of recipes and then turn them into a cookbook, or it can create interactive flashcards based on a series of video lectures.
You’ll see some of these improvements across Google’s suite of products, including the Gemini app, where you can create more “full-featured” programs inside the built-in workspace, Canvas. The advanced AI models will also enable “generative interfaces”, a tool that Google is testing in Gemini Labs, which allows the Gemini 3 Pro to create a visual, magazine-style format with images that you can browse, or create a dynamic layout with a custom user interface tailored to your prompt.
Gemini 3 Pro in AI mode – the AI-powered Google search feature – will present you with visual elements like images, tables, grids, and simulations based on your query. It’s also able to perform more searches using an advanced version of Google’s “query fan-out technology”, which now not only breaks queries into chunks it can search on your behalf, but also “better understands intent to help find new content that was previously missed,” according to Google’s announcement.
Google also isn’t so subtly poking fun at OpenAI, describing the Gemini 3 Pro as less susceptible to the kind of empty chatter favored by ChatGPIT. Doshi says you’ll see a “noticeable” change in responses to the Gemini 3 Pro, which Google describes as “smart, concise and direct, trading business clichés and sycophancy for real insights – telling you what you need to hear, not just what you want to hear.” The company says it also shows “less sycophancy”, an issue that OpenAI had to address with ChatGPT earlier this year.
According to Google, with these improvements, Gemini 3 Pro comes with improved reasoning and agentic capabilities, allowing it to complete more complex tasks and “reliably plan ahead over longer horizons.” The AI model is powering an experimental Gemini Agent feature that can perform tasks on your behalf inside the Gemini app, like reviewing and organizing emails, or researching and booking travel.
The Gemini 3 Pro now tops the leaderboard of LMArena, a popular platform used for benchmarking AI models. Deep Think mode further enhances the model’s reasoning capabilities, although it is currently only available to security testers.
Gemini 3 Pro is available inside the Gemini app for everyone starting today, while Google AI Pro and Ultra customers in the US can try the Gemini Agent in the Gemini app, as well as try Gemini 3 Pro inside AI mode by selecting “Thinking” from the Model dropdown.