Gemini 3 Deep Think mode is now available to Google AI Ultra subscribers through the Gemini app. This feature significantly enhances the AI's reasoning abilities, specifically for tackling challenging math, science, and logic problems. The improvements position Gemini 3 Deep Think as a leader in performance metrics, achieving impressive scores on benchmarks such as Humanity’s Last Exam with 41.0% accuracy, and ARC-AGI-2, where it reached 45.1% with code execution. The mode employs advanced parallel reasoning, allowing it to evaluate multiple hypotheses at once. This builds on the success of the Gemini 2.5 Deep Think variants, which recently earned gold medals at prestigious competitions like the International Mathematical Olympiad and the International Collegiate Programming Contest World Finals. Users can access this new mode by selecting “Deep Think” in the prompt bar and choosing Gemini 3 Pro in the model dropdown. For those interested in pushing the boundaries of AI reasoning, Gemini 3 Deep Think offers substantial advancements compared to previous iterations. The combination of its high benchmark scores and advanced reasoning techniques suggests a significant leap forward in AI capabilities.