Google is releasing the high-performing Deep Think AI to select researchers, supporting advanced reasoning tests and future optimization in complex math tasks.

The Gemini 2.5 Deep Think model that achieved the gold-medal standard will be shared with a small group of mathematicians and academics. The intention is that this model will be used to advance their work, and it is hoped the lessons will provide feedback for improvements.

Deep Think is based on the same foundation as Gemini 2.5 Pro, but it increases the "thinking time" with greater parallel analysis. According to Google, Deep Think explores multiple approaches to a problem, even revisiting and remixing the various hypotheses it generates. This process helps it create a higher-quality output.

The company says Gemini 2.5 Deep Think achieved new performance records—for AI, at least—on common knowledge and reasoning tests.

Google released its first publicly available "multi-agent" AI system, which uses more computational resources, but produces better answers.

Google AI Ultra subscribers now have access to Deep Think, Google’s most advanced reasoning model yet for Gemini 2.5. The new model was first unveiled at Google I/O, and after feedback from "early trusted testers,

Google says the model outperformed several of its competitors in key AI benchmark tests and a variation of the model even achieved the gold-medal standard at this year’s International Mathematical Olympiad (IMO),

A version of the AI model achieved gold-medal standard at the International Mathematical Olympiad earlier this year.

Google DeepMind, Alphabet Inc.’s artificial intelligence research arm, today announced the rollout of Gemini 2.5 Deep Think, a new creative problem-solving AI model. The company stated the model is capable of addressing complex problems by considering multiple ideas simultaneously and choosing the best answer.