
UnifoLM (Unified Robot Large Model) is the name used by Unitree Robotics for its large model layer and robotic intelligence architectures. On the official Unitree G1 page, UnifoLM is presented as a component for building a "new era of intelligence" for embodied AI robots.
Unitree's public materials suggest that UnifoLM is not a single, simple SDK but a family of related models and frameworks. Unitree's open-source page lists, for example, UnifoLM-VLA-0 as a Vision-Language-Action model for general humanoid manipulation, while the G1-D page presents UnifoLM-WMA-0 as a World-Model-Action architecture for general-purpose robot learning.
For this reason, the "UnifoLM" entry is best treated in the catalog as the overarching Unitree embodied AI family or platform, with more specific variants β such as UnifoLM-VLA-0 and UnifoLM-WMA-0 β potentially added as separate, more detailed entries in the future.
Perception
No additional description for this role.
Motion Planning
No additional description for this role.