The Handoff
A generalist control model moves the arm; a specialist grasp planner decides how to pick the object up.
Watch the live demoGR00T1 — a generalist model for many robots.
But it learned standard two-finger grippers — not multi-finger hands or suction cups.
A gripper can't hold a big object.
A standard two-finger gripper can't wrap a large object like an apple — it never forms a stable hold, so the lift fails.
A bigger hand can hold it — GR00T can't drive it.
The hardware can hold it. The generalist just never learned this hand — so control breaks the moment we change it.
Models that focus only on the grasp.
TRO-Grasp2
A planner for a multi-finger Barrett hand — where to place each finger for a stable hold.
Suction-Grasp3
A planner for a vacuum cup. Picks the best spot and angle to seal onto the object and lift it.
Each answers one question: given this object's shape, how should the hand make contact?
Compose,
don't retrain.
Keep the generalist model unchanged. Bring in a specialist just for the grasp, then hand control back — a modular combination of generalist and specialist, with no extra training.
The model and the planner take turns.
Reach
The control model moves the arm to the object.
Grasp
Control passes to the planner, which sets the hand and secures the object.
Carry & place
Control returns to the model, which lifts, moves, and releases.
A small controller decides when to pass control — the model never has to learn the grasp itself.
One interface. Any brain, any hand.
The link is just a standard message, so either side can be replaced on its own. TRO-Grasp2 and Suction-Grasp3 are only two instances — the same recipe accepts a new control model or a new end-effector, with no extra training.
Control model
interface
Grasp specialist
Combine generalist & specialist
Same recipe — a frozen generalist plus a grasp specialist — across two robots and two different end-effectors. No retraining either time.
Where the handoff is new.
| Approach | Generalist, language-driven |
Specialist grasp (dexterous + suction) |
No retraining |
Many robot arms |
Any end-effector |
|---|---|---|---|---|---|
|
End-to-end VLAs
RT-2 · OpenVLA · π0 · GR00T
|
|||||
|
Dexterous-grasp VLA
DexGraspVLA
|
|||||
|
Grasp synthesizers
AnyGrasp · T(R,O) · SuctionNet
|
|||||
|
Modular VLM agents
MAESTRO · VoxPoser
|
|||||
|
The Handoff (ours)
GR00T × TRO-Grasp / Suction
|
Each prior approach covers part of the matrix; composing frozen modules over one standard interface is what covers all of it — generalist control and specialist grasps, across robots and end-effectors, with no retraining.