Dynamic3D · Sweep simple¶
Agent runs
Each mode of this task runs once per run. Est. cost is tokens at list price (data/prices.yml), never a bill; see the KinDER runs for every task of a run.
Codex CLI 0.157–0.159 + GPT-6 Luna, reasoning medium (OpenRouter)default run
| Trial | Started | Agent time | Model requests | Tokens in / out | Est. cost | Grade |
|---|---|---|---|---|---|---|
| U✓ 56m | 10-04 06:17 | 56m | — | 13.9M / 185k | $0.284 | deterministic 1, grader_error 0, missing_trajectory 0, n_actions 914, video_rendered 1 |
| L✗ 25m | 10-04 06:33 | 25m | — | 16.7M / 117k | $0.244 | deterministic 1, grader_error 0, live_success 0, missing_log 0, n_actions 1000, replay_success 0, video_rendered 1 |
1 other run(s) of KinDER
Codex CLI 0.159.2 + GPT-6.1 Sol, reasoning medium (OpenRouter)
| Trial | Started | Agent time | Model requests | Tokens in / out | Est. cost | Billed | Grade |
|---|---|---|---|---|---|---|---|
| U✓ 27m | 10-01 06:03 | 27m | 90 | 7.3M / 32k | $1.29 | $1.35 | deterministic 1, grader_error 0, missing_trajectory 0, n_actions 982, video_rendered 1 |
| L✓ 22m | 10-01 02:55 | 22m | 60 | 3.4M / 24k | $0.756 | $0.803 | deterministic 1, grader_error 0, live_success 1, missing_log 0, n_actions 980, replay_success 1, video_rendered 1 |
Task instruction (upstream)
Get the ten small dark-red cubes scattered on the floor into the translucent grey box marked on the floor at the end of the kitchen island (the free-standing white-topped counter) nearest the robot; a wiper stands on the floor next to the cubes. The task succeeds the moment all ten cubes' centres are inside that box.
What KinDER states about this task
| Success criteria | 1. KinDER's own goal check (the environment's terminated) fires within the episode: at most 1000 control steps (100 s at 10 Hz); no partial credit2. unlimited: both fresh-process replays of the handed-in trajectory reach it and end in the same state 3. limited: the one episode reaches it (no reset in our runs); the service records the episode, and it replays to the same state |
| Env Id | kinder/SweepSimple3D-o10-sweep_the_blocks_to_the_left_side_of_the_kitchen_island-v0 |
| Robot | TidyBot++: holonomic base, Kinova Gen3 7-DoF arm, Robotiq 2F-85 gripper |
| Category | Dynamic3D (MuJoCo) |
| Instance | seed 0: the scene KinDER builds from it, pinned by its digest |
| Holding Still | does not reach the goal (checked by running an episode of zero actions) |
| Action Dim | 11 |
| Actions | base pos and yaw (3), arm joints (7), gripper pos (1); each base and joint delta at most 0.1 a step |
| State Dim | 243 |
| Limited Mode | base RGB and wrist RGB-D cameras (480×640), base odometry, the arm's joint angles and velocities, the fingers' position. Not where any object or the goal region is, not the goal check before the episode ends |
| Agent Budget | 3600 s of wall clock per mode |
From https://github.com/Princeton-Robot-Planning-and-Learning/kindergarden @ 5b2dbac, as defined in our task definitions @ ae0d92a50.
Tags¶
Why this task is interesting¶
Ten 2 cm cubes on the floor go into a box marked at the end of the kitchen island, 0.6 to 1.1 m away. The base passes over the cubes without moving them, so only the gripper or a standing T-shaped wiper can.
Capability notes¶
Not yet written.
Oracle demo review¶
No demo.
Discussion¶
Passed by carrying each cube to the box rather than sweeping it: GPT-6.1 Sol in both modes, after the wiper rode over the cubes in unlimited mode, and GPT-6 Luna (2026-10-04) in unlimited mode, after moving the wiper out of the cubes' way (goal at step 914). (@williamzhangNU)