New Delhi: Google has officially announced a major update to its reasoning mode, Gemini 3 Deep Think. This latest version is especially designed to handle complex scientific and engineering problems while also supporting practical, real-world use cases. This latest upgrade is now rolling out to Google AI Ultra subscribers, and it will also be available to selected researchers and enterprises via the Gemini API.
The actual aim is to bring Deep Think directly to professionals who need the advanced reasoning tools in their research and development work.
One of the latest features in this recent update enables users to convert a simple sketch into a 3D printable model. Deep Think will generally analyze the drawing, build the required 3D structure, and generate a file that can be used for 3D printing. Google has also stated that the feature highlights how the system can connect the reasoning with hands-on engineering tasks. The users can access this by selecting the Deep Think option from the tools menu in the Gemini application.
Google has also said that this upgraded Deep Think will continue to improve on academic benchmarks, because it has been used for research-level mathematics exploration and competitive coding tasks. This model has achieved gold medal-level results on the written sections of the 2025 International Physics Olympiad and Chemistry Olympiad.
It has also shown strong performance in advanced theoretical physics, as well as in broader areas such as chemistry and physics.
