TABLETOP / LAB
HOSTED CONSOLE| GR00T N1.6 × FOURIER GR-1

EXPERIMENT 001 / LANGUAGE → ACTION

Give the robot
something to do.

A tabletop, two drinks, an open box.
Type an instruction and watch a humanoid
policy work through it in simulation.

SIMULATION FEED
DRAG TO ORBIT
+
GR-1 IS WAITING FOR A TASK

Your first run loads the policy on Modal.
The GPU sleeps after 10 minutes without a request.

NVIDIA GR00T / ROBOT VIEW + POLICY CAMERA
SIMULATION0.0 s
MODEL INFERENCE— ms / chunk
STEPS0 / 400
RESULTAWAITING TASK
↳Enter a command to begin.
01 / TASK INPUT3B POLICY

Instructions go directly to GR00T. Each run resets the scene; the same seed lets you compare commands.

02 / SCENE OBJECTS

●Soda canON TABLE
▯Water bottleON TABLE
□Cardboard boxRECEPTACLE

The run log00

Each attempt leaves a recording and a result here.