Dynamic3D · Dynamo¶
Excluded from the benchmark
a fixture, not scored: it checks the pipeline end to end (drive into a floor square past three chairs)
Greyed and last in the task list, left out of the benchmark's numbers; its demo and agent runs stay (excluded: in state/tasks/kinder.yml).
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✓ 4m | 10-01 06:13 | 4m | — | 328k / 8k | $0.0095 | deterministic 1, grader_error 0, missing_trajectory 0, n_actions 577, video_rendered 1 |
| L✓ 7m | 09-30 03:40 | 7m | — | 2.3M / 29k | $0.044 | deterministic 1, grader_error 0, live_success 1, n_actions 724, replay_success 1, 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✓ 1m | 10-01 05:51 | 1m | 9 | 168k / 2k | $0.086 | $0.099 | deterministic 1, grader_error 0, missing_trajectory 0, n_actions 50, video_rendered 1 |
| L✓ 4m | 10-01 02:41 | 4m | 31 | 749k / 5k | $0.190 | $0.207 | deterministic 1, grader_error 0, live_success 1, missing_log 0, n_actions 183, replay_success 1, video_rendered 1 |
Task instruction (upstream)
Drive the robot into the translucent green square marked on the floor; three chairs stand in the way and can be pushed. The task succeeds the moment the centre of the robot's base is inside the square. (The red, blue and yellow squares only mark where the chairs started.)
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/Dynamo3D-o3-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 | 70 |
| 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¶
A fixture: drive the base into a square marked on the floor past three chairs, which can be pushed. Without a camera in the room, the base camera sees only the floor in front of it, so the square has to be found first. It checks the whole pipeline end to end and is not scored.
Capability notes¶
Not yet written.
Oracle demo review¶
No demo.
Discussion¶
Works as the fixture: both modes pass, limited mode in its first episode by locating the square with wrist depth. (@williamzhangNU)