Microsoft has unveiled Copilot 3D, an innovative AI-powered feature that empowers users to convert 2D images into fully rendered 3D models, no modeling experience required. Available via Copilot Labs and accessible globally to users with a personal Microsoft account, this experimental tool is making 3D creation fast, intuitive, and widely accessible.
All it takes is a simple JPG or PNG image under 10MB, preferably with clear background separation, and Copilot 3D automatically generates a downloadable GLB-format model. These creations are stored in the "My Creations" archive for up to 28 days, giving users time to preview or export for use in games, animation, 3D printing, VR/AR, and other creative workflows.
How to create 3D models from JPG?
Head to the Copilot 3D page and register using a Microsoft, Apple, or Google account.
Upload an image, and click on Generate to create the 3D model of the image.
As mentioned, it is ideal to upload an image with clear background separation.
Download the 3D model in GLB format.
The tool is free to use for all, and the only thing you need to do is register.
Hands-on tests reveal stark contrasts in output: static inanimate objects, like furniture and bananas, translate beautifully into 3D, while complex subjects such as electronics or animals produce less reliable results. Microsoft’s safeguards also prevent users from converting images of public figures or copyrighted content, with suspended access for violations.
Positioned as a creative equalizer, Copilot 3D taps into the power of GPT-5, later integrated into Copilot, making AI-assisted visual experimentation smoother than ever.
