Advanced AI Technologies

Reka AI previews multimodal model for robot control

Reka AI previewed Rho-1, a 19-billion-parameter multimodal model spanning text, images, video, and robot control.

Reka AI released a research preview of Rho-1, a 19-billion-parameter model designed to handle text, images, video, and robot control within one system. The model therefore targets both conventional multimodal understanding and physical-machine interaction.

Its inclusion of robot control distinguishes the preview from models focused only on text or media generation. A system combining language, visual information, video, and control could support robotics applications in which an agent must interpret its environment and translate instructions into actions.

The research preview indicates that the technology is not necessarily a mature commercial product. Its significance lies in the direction of model development: a single model is being designed to operate across digital content and embodied systems.

For advanced AI developers, multimodal robot-control models could support future automation in logistics, manufacturing, and other environments. The available report does not state the model’s benchmark performance, hardware requirements, or production customers.

Original source: AI Daily Post

Related RAQI solutions

In-depth analysis on RAQI Insight

Work with RAQI