Google's Gemini Deep Think explores ideas simultaneously

Google launched Gemini 2.5 Deep Think, an advanced AI reasoning model that explores multiple ideas simultaneously to find the best answers. This new "multi-agent" system, available to Ultra subscribers, uses more computational power but delivers improved results, outperforming competitors on benchmark tests like Humanity's Last Exam and LiveCodeBench6. The model, first shown in May, is Google's first publicly available multi-agent AI and has demonstrated success in complex tasks, including winning a gold medal at the International Math Olympiad.