Google has officially launched Gemini 2.5 Deep Think mode for AI Ultra subscribers following its preview at I/O 2025. This advanced feature utilizes parallel thinking techniques, allowing Gemini to generate and evaluate multiple ideas simultaneously, enhancing its problem-solving capabilities over time. Deep Think benefits from extended reasoning paths, allowing it to creatively tackle complex issues. With significant improvements in coding and reasoning—scoring 87.6% on LiveCodeBench—Deep Think is optimized for iterative development, making it adept at enhancing both the aesthetics and functionality of web development tasks. It excels in solving tough coding challenges and complex scientific inquiries, potentially expediting discoveries in fields like mathematics and science. Priced at $249.99 per month, users can access Deep Think through Gemini 2.5 Pro. Google plans to introduce versions of Deep Think to developers and enterprises via the Gemini API soon. This rollout marks a pivotal enhancement in AI-driven problem-solving tools.
Source link
Share
Read more