Roblox has unveiled Cube 3D, an open-source AI model designed to generate 3D objects, marking its first step toward AI-driven content creation. The beta version of Cube 3D’s mesh generation feature will be available this week in Roblox Studio and as an in-experience Lua API.
“Cube will underpin many of the AI tools we’ll develop in the years to come, including highly complex scene-generation tools. It will ultimately be a multimodal model, trained on text, images, video, and other types of input—and will integrate with our existing AI creation tools,” the company stated.
What is Cube 3D?
Cube 3D is an AI-powered foundational model that can generate 3D models and environments from text inputs. Future versions are expected to incorporate image-based inputs as well. Unlike conventional methods that rely on 2D image reconstruction, Cube 3D is trained on native 3D data, which the company claims enhances the functionality and realism of its outputs.
To encourage customization and broader use, Roblox is releasing a version of Cube 3D for developers and researchers. The model’s Mesh Generation API, accessible in Roblox Studio and Lua API, enables users to create 3D objects from simple text prompts.
How Does Cube 3D Work?
Cube 3D leverages a technique called 3D tokenization, which converts 3D shapes into tokens, similar to how language models break down text. This allows the AI to predict and generate subsequent shape tokens, eventually forming complete 3D objects and scenes.
The model is based on autoregressive transformers, a widely used architecture for text generation. Thanks to its scalable design, Cube 3D can process multiple input types, including text, images, and video, positioning it as a multimodal AI model.
Currently, Cube 3D enables the creation of 3D assets such as furniture, vehicles, clothing, weapons, and animals using simple text prompts. Roblox has also indicated that future updates will introduce features like scene creation, object interactivity, and dynamic content generation.
