Google Gemini Enables 3D Printing

- The sketch-to-3D feature is part of a major upgrade to Gemini 3's "Deep Think" reasoning mode, which was developed in partnership with scientists and researchers to solve complex problems where data may be incomplete. This specialized mode is now accessible to Google AI Ultra subscribers through the Gemini app and to select researchers and enterprises via the Gemini API. - While Google hasn't released a detailed technical paper on this specific feature, the underlying technology for this type of 2D-to-3D generation likely involves a combination of diffusion models to interpret the sketch and Neural Radiance Fields (NeRFs) to construct the 3D representation. Google Research has previously detailed using view-conditioned diffusion models and NeRFs to create 3D product visualizations from 2D images. - The upgraded Deep Think model, which powers this feature, has demonstrated state-of-the-art performance on several advanced reasoning benchmarks. It achieved an 84.6% score on ARC-AGI-2, a test for solving novel problems, and an Elo of 3455 on the Codeforces competitive programming benchmark. - This functionality moves beyond theoretical applications by integrating deep reasoning with practical engineering tasks, allowing users to go directly from a conceptual drawing to a physical object file. The system analyzes the drawing, models the geometry, and generates a file ready for printing. - The release of this feature positions Google in a competitive landscape of generative 3D modeling. Other notable models include OpenAI's Shap-E, which also generates 3D assets from text prompts, and various research models like Sketch2NeRF that explore sketch-guided 3D generation. - Access to this advanced feature is through the "Deep Think" option within the tools menu of the Gemini app for Google AI Ultra subscribers. Researchers and enterprises can also apply for early access to integrate these capabilities into their own workflows via the Gemini API.