Google has introduced its Gemini 2.5 Deep Think AI model to Google AI Ultra subscribers, bringing advanced reasoning capabilities and the ability to solve complex problems using parallel thinking techniques.
The AI model is designed for complex math and coding tasks and provides more detailed, creative, and thoughtful responses to prompts. In benchmark tests, Gemini 2.5 Deep Think outperformed competing models, including xAI’s Grok 4 and OpenAI’s o3, scoring 34.8% on the Humanity’s Last Exam test.
The model also achieved a gold-medal standard at the International Mathematical Olympiad (IMO) earlier in July. Google will offer this version to select “trusted testers”, including mathematicians and academics. The AI model is part of the Gemini app and will be available starting Friday, August 1, for $250 per month.
Source: https://mashable.com/article/deep-think-google-gemini-app-available