Google launches Gemini 3 Deep Think with advanced reasoning: Who can use it

Google has unveiled its latest advancement in artificial intelligence, the Gemini 3 Deep Think, which is now available to Google AI Ultra subscribers. This innovative feature, announced on Thursday, marks a significant enhancement in problem-solving capabilities, especially for users dealing with complex mathematics, science, and logic tasks. The Deep Think mode is touted as a major upgrade from its predecessors, allowing the AI to process multiple reasoning pathways simultaneously. This breakthrough greatly enhances its ability to tackle challenges that would typically strain even the most sophisticated AI models. According to Google, Gemini 3 Deep Think has excelled in some of the most rigorous evaluation tests, achieving a notable 41.0% score on Humanity’s Last Exam without tools, and an impressive 45.1% on ARC-AGI-2 when code execution is involved. This new release builds upon the success of the Gemini 2.5 Deep Think versions, which had previously demonstrated gold-medal level performance in prestigious competitions like the International Mathematical Olympiad and the ICPC World Finals. Ultra subscribers can immediately access the new Deep Think mode, which can be activated through the Gemini app by selecting "Deep Think" from the prompt bar and choosing the Gemini 3 Pro model. In an announcement on X, the official Gemini account described the feature as the most advanced reasoning mode yet, capable of exploring multiple hypotheses at once to deliver more sophisticated outputs. With the launch of Gemini 3 Deep Think, Google is making significant strides in advanced computational reasoning, providing robust support for users facing intricate academic, technical, or analytical challenges. This release follows the introduction of Gemini 3 last month, which was positioned as the new benchmark in the AI field, surpassing the capabilities of the previously dominant Gemini 2.5 Pro model.