CAD-MLLM is the first unified system capable of generating parametric CAD models from multiple input types (text, images, point clouds, or combinations), revolutionizing engineering design workflows.

• Leverages command sequences of CAD models with large language models
• Processes multiple input modalities simultaneously for flexible design generation
• Enables users to express design intent through their preferred communication method
• Streamlines the engineering design process by reducing the gap between concept and technical implementation

For engineering teams, this breakthrough means faster prototyping, more intuitive design interfaces, and the ability to transform rough concepts into precise CAD models through natural communication channels.

CAD-MLLM: Unifying Multimodality-Conditioned CAD Generation With MLLM