Put the bowl on top of the cabinet

libero-goal-4LIBEROFranka PandaManipulationmedium

Reference solution, 81 of 300 steps.

Instruction

A Franka Panda arm with a parallel-jaw gripper works at a table in a LIBERO scene (MuJoCo / robosuite; LIBERO suite libero_goal, task put_the_bowl_on_top_of_the_cabinet). LIBERO-Goal: one fixed kitchen scene; the instruction sets the goal.

Goal (LIBERO's language instruction): put the bowl on top of the cabinet.

The episode ends as solved the moment LIBERO's own success check (the task's goal predicates) passes; you do not need to call robo done after that. If you finish without the check passing, call robo done (or robo give-up).

This robot

The action has 7 numbers, not 4: robo act DX DY DZ DROLL DPITCH DYAW GRIPPER, each in [-1, 1]. It is LIBERO's delta end-effector command (robosuite's OSC_POSE controller, 20 steps per second). Holding DX/DY/DZ at 1.0 moves the hand about 1 cm per step; DROLL/DPITCH/DYAW rotate the hand (1.0 asks for 0.5 rad; usually leave them at 0). GRIPPER -1 opens, +1 closes.
robo move-to X Y Z [--grip G] and robo grip G are available (they keep the hand's orientation).
World frame, metres: +x points forward, away from the robot's base; +y is to the robot's left; +z is up. The table surface is at about z = 0.90.
robo observe fields: hand_pos (end-effector position), eef_quat (orientation quaternion x, y, z, w) and eef_axis_angle, gripper_qpos (the two finger joints), gripper_open (finger gap in metres, about 0.08 when fully open), and objects: the position and quaternion of every object in the scene, under LIBERO's names (e.g. akita_black_bowl_1, plate_1; ..._1, ..._2 number identical objects); fixtures: the position and quaternion of each piece of furniture (cabinet, stove, microwave, rack, shelf, ...); articulated: each movable part of a fixture (drawer slide, door hinge, stove knob), keyed by its joint name (e.g. wooden_cabinet_1_top_level is the cabinet's top drawer), with qpos (joint position), range (joint limits), LIBERO's thresholds for that fixture (open_ranges / close_ranges, or turnon_ranges / turnoff_ranges for a stove knob: the joint positions at which LIBERO counts the part as open / closed or on / off) and box_min / box_max (the world-frame bounding box of the moving part).
robo observe --image saves the front camera (agentview); add --camera robot0_eye_in_hand for the wrist camera. Both are 256x256.
The step budget is 300 steps (LeRobot's LIBERO evaluation budget for this suite).

How the robot is controlled and scored

You are controlling a simulated robot. Read the task below, then solve it by running the robo command in your shell (start with robo info and robo observe). Keep going until the task is done, then call robo done once. Do not stop to ask questions; there is no human to answer.

How to control the robot

You are the robot's policy. You act only through the robo command in your shell. There is no other way to move the robot, and you cannot read or change the simulator, the scoring, or other files to succeed; the episode server judges the final physical state itself.

robo info                         # action space, available skills, step budget
robo observe                      # robot and object state as numbers
robo observe --image              # also saves a camera image and prints its path (open it to look)
robo act DX DY DZ GRIP [--repeat N]   # low-level action, applied N times (N <= 50)
robo move-to X Y Z [--grip G]     # skill: move the gripper toward a point (if enabled for this task)
robo grip G [--steps N]           # skill: hold position and set the gripper (+1 close, -1 open)
robo done "short summary"         # end the episode and ask for scoring
robo give-up "reason"             # end the episode without claiming success

Positions are in metres in the world frame (x, y on the table plane, z up).
The episode has a fixed step budget (see robo info); every simulated step counts, including skills.
Unless the task says otherwise, success is judged about 10 steps after you call robo done, with the robot holding still, so the goal must still be true when the robot stops.
Work in small steps and re-observe after each motion. Call robo done exactly once when finished.

Run this task

ROBOUSE_ORACLE_TOKEN=$(openssl rand -hex 16) \
  bench eval run \
    -d benchflow/robouse-core@0.3 \
    --registry https://robouse.ai/hub/registry.json \
    --agent oracle \
    --include libero-goal-4

Pinned to robohub commit 954472c78299. The verifier and the reference solution are not published.

Filter models

					Trial
Kimi K3 · Claude Code	Solved	298 / 300	6 min	Sep 28, 2026	Trial Compare
Kimi K3 · mini-swe-agent	Solved	81 / 300	4 min	Sep 28, 2026	Trial Compare
GLM-5.3 · Claude Code	Solved	193 / 300	4 min	Sep 27, 2026	Trial Compare
pi0.5-LIBERO · VLA	Solved	76 / 300	32 s	Sep 30, 2026	Trial Compare
GR00T N1.7-LIBERO · VLA	Solved	83 / 300	23 s	Sep 30, 2026	Trial Compare
EVO1-LIBERO · VLA	Solved	79 / 300	54 s	Sep 30, 2026	Trial Compare
VLA-JEPA-LIBERO · VLA	Solved	78 / 300	42 s	Sep 30, 2026	Trial Compare
pi0.5-LIBERO (LeRobot) · VLA	Solved	78 / 300	18 s	Sep 30, 2026	Trial Compare
pi0-LIBERO · VLA	Solved	89 / 300	31 s	Sep 30, 2026	Trial Compare
pi0-FAST-LIBERO (LeRobot) · VLA	Solved	88 / 300	24 s	Sep 30, 2026	Trial Compare
pi0-LIBERO (LeRobot) · VLA	Solved	83 / 300	31 s	Sep 30, 2026	Trial Compare
MolmoAct2-LIBERO · VLA	Solved	81 / 300	52 s	Sep 27, 2026	Trial Compare
SmolVLA-LIBERO · VLA	Solved	82 / 300	22 s	Sep 30, 2026	Trial Compare
GPT-6 Astra · Codex	Not solved	290 / 300	2 min	Sep 27, 2026	Trial Compare
GLM-5.3 · mini-swe-agent	Not solved	286 / 300	6 min	Sep 27, 2026	Trial Compare
GLM-5.3 · Codex	Not solved	271 / 300	2 min	Sep 27, 2026	Trial Compare

These trials ran on version 0.1, now deprecated (superseded by 0.3).