These are top 10 stocks traded on the Robinhood UK platform in July
Investing.com -- Google (NASDAQ:GOOGL) has launched Deep Think in the Gemini app for Google AI Ultra subscribers, bringing advanced problem-solving capabilities to users. The feature represents a significant improvement over the version announced at I/O, according to the company.
Deep Think is a variation of the model that achieved gold-medal standard at this year’s International Mathematical Olympiad (IMO). While the competition version takes hours to process complex math problems, the consumer release is faster for everyday use while still reaching Bronze-level performance on the 2025 IMO benchmark in internal evaluations.
Google is also sharing the official version of the Gemini 2.5 Deep Think model with a select group of mathematicians and academics to gather feedback for future improvements.
The technology works by extending Gemini’s "thinking time" and using parallel thinking techniques that allow it to generate and consider multiple ideas simultaneously. Google has developed reinforcement learning techniques that encourage the model to use these extended reasoning paths, enabling it to become a more effective problem-solver.
Deep Think excels at tasks requiring creativity and strategic planning, including iterative development, scientific discovery, and algorithmic coding. The company reports that the model achieves state-of-the-art performance on benchmarks like LiveCodeBench V6 and Humanity’s Last Exam when compared to other models without tool use.
Google AI Ultra subscribers can access Deep Think in the Gemini app by toggling the feature in the prompt bar when selecting 2.5 Pro in the model dropdown. The feature works with tools such as code execution and Google Search, and can produce longer responses.
The company plans to release Deep Think to trusted testers via the Gemini API in the coming weeks to better understand its potential for developer and enterprise applications.
This article was generated with the support of AI and reviewed by an editor. For more information see our T&C.