Galbot · 银河通用
Embodied / robotics model — runs on robot hardware, not a token API. No per-1M pricing. For API-priced LLMs see the model catalog.
| Creator | Galbot (银河通用) |
| Architecture | Vision-Language-Action (VLA) |
| Embodiment | Mobile manipulation (pick-and-place, navigation) |
Galbot takes an intelligence-first approach with a family of narrow-remit VLA models powering its G1 robot rather than one generalist model: GraspVLA for dexterous grasping, TrackVLA for navigation, and GroceryVLA for retail pick-and-place. The split illustrates a real engineering trade-off — fully-autonomous general manipulation is hard, so Galbot ships reliable specialists. Models are tied closely to Galbot's own hardware + deployment rather than open weights.
Retail and warehouse pick-and-place automation; reference point for the specialist-vs-generalist VLA debate.
Your brand here — Reach our audience of professional directory owners and boost your sales.