Google officially begins the release of Gemini 3 today, introducing a model series the company claims is its "most intelligent" and "factually accurate" to date, The Verge reported.
This launch grants universal access to the flagship Gemini 3 Pro via the Gemini app immediately, positioning Google to compete aggressively with OpenAI following the rocky debut of GPT-5.
According to The Verge, Tulsee Doshi of Google DeepMind stated that the new model moves the company closer to making information "universally accessible and useful".
"I think the one really big step in that direction is to step out of the paradigm of just text responses and to give you a much richer, more complete view of what you can actually see," Doshi told The Verge.
The system is "natively multimodal," meaning it processes images, text, and audio simultaneously, allowing it to convert recipe photos into cookbooks, The Verge noted.
Users can build "full-featured" programs in the Canvas workspace, while "generative interfaces," in Gemini Labs create dynamic visual layouts tailored to prompts.
The Verge reported that Gemini 3 Pro in AI Mode uses an upgraded "query fan-out technique," to understand intent and "find new content that it may have previously missed,".
Google describes the model as "smart, concise and direct, trading cliche and flattery for genuine insight – telling you what you need to hear, not just what you want to hear."
The company claims the model displays "reduced sycophancy," a direct improvement over issues found in earlier chatbots like ChatGPT, The Verge highlighted.
Enhanced agentic capabilities allow the AI to "reliably plan ahead over longer horizons," enabling experimental features that can book travel or organize emails.
Gemini 3 Pro currently leads the LMArena benchmarking leaderboard, with a more advanced Deep Think mode available exclusively to safety testers, The Verge stated.
The Verge confirmed that Gemini 3 Pro is available in the Gemini app today, while US subscribers to Google AI Pro and Ultra can access it in AI Mode by selecting "Thinking".
