Roblox announced the launch of the first version of its generative 3D model called Cube, which enables creators to generate three-dimensional objects using generative AI. The Cube model is already available as open access, allowing anyone outside the platform to use it for their own needs. This release was announced last year at Roblox’s annual developer conference and will be showcased today at the Game Developers Conference (GDC).
The first version of Cube includes a 3D mesh generation tool, currently available in beta. With a single text prompt, creators can generate 3D objects, such as “an orange race car with black stripes.” In addition, developers can customize these objects in Roblox Studio. The open version of Cube 3D allows users to create their own plugins or train the model on their own datasets.
Roblox also announced three additional AI-based tools — text generation, text-to-speech, and speech-to-text. These features will become available in the coming months. The text generation tool will allow adding AI-powered text features to games, such as conversations with interactive characters. The text-to-speech feature will enable voiceovers, while speech-to-text will let players use voice commands.
In the future, Roblox plans to launch mesh generation for more complex objects and scene generation, which will make it possible, for example, to create entire forest scenes with changing leaf colors. The company sees Cube as a way to speed up workflows and give indie developers the ability to tackle larger projects. However, the use of AI raises concerns in the industry, as according to research by CVL Economics, by 2026, 13.4% of jobs in gaming could be replaced by AI.
