Google has launched Deep Think, its most advanced reasoning model to date for the Gemini platform, now available exclusively to Google AI Ultra subscribers for $250 per month. The model was first unveiled at Google I/O 2025 and has rolled out broadly following feedback from early trusted testers.
The public Deep Think variant is adapted from a version that achieved a gold-medal performance standard at the International Mathematical Olympiad. According to internal testing, this release version delivers bronze-medal capabilities while being optimized for daily usage speed and efficiency. The original gold-level model remains under evaluation with mathematicians and academics.
Deep Think utilizes parallel thinking techniques—processing multiple ideas simultaneously—while supporting multimodal inputs including text, images, and audio. In benchmark tests like Humanity’s Last Exam and LiveCodeBench, it outperformed competitors including OpenAI’s o3, xAI’s Grok 4, and Google’s own Gemini 2.5 Pro.
Access requires a premium AI Ultra subscription. Users can interact with Deep Think via the Gemini app, though Google imposes undisclosed daily prompt limits. The company positions this release as a major advancement in AI reasoning capabilities for specialized users.
