Clean Up Toys¶
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 RoboLab runs for every task of a run.
Codex + GPT-6 Luna, reasoning xhigh (ChatGPT login)default run
| Trial | Started | Agent time | Model requests | Tokens in / out | Est. cost | Grade |
|---|---|---|---|---|---|---|
| U✗ 1h 30m 78% ↻ | 09-27 03:51 | 1h 30m | 174 | 22.1M / 83k | $0.288 | deterministic 1, grader_error 0, instance_exact 1, instance_ok 1, n_actions 6523, video_rendered 1 |
| L✗ 13m | 09-29 10:20 | 13m | 34 | 627k / 8k | $0.014 | deterministic 0, grader_error 0, live_success 0, n_actions 0, replay_success 0, video_rendered 0 |
Task instruction (upstream)
Clean up all the smaller toys and leave the birdhouse on the table
No oracle video published upstream — this is the scene it starts from.
What RoboLab states about this task
| Success predicate | object_in_container |
| Predicate arguments | object: ['rubiks_cube', 'rubiks_cube_1', 'rubiks_cube_2', 'lizard_figurine', 'yellow_block', 'red_block', 'green_block', 'blue_block', 'lizard_figurine_01'] · container: grey_bin · logical: all · require_gripper_detached: True |
| Scored subtasks | 1 |
| Subtask predicates | pick_and_place |
| Objects | rubiks_cube rubiks_cube_1 rubiks_cube_2 grey_bin lizard_figurine birdhouse yellow_block red_block green_block blue_block lizard_figurine_01 |
| Upstream attributes | sorting semantics |
| Upstream difficulty | simple |
| Episode budget | 300 s |
| Other wordings | vague — Clean up everything except the birdhouse specific — Pick up every toy object (rubiks cube, lizard figurines, colored blocks) from the table and place them all into the bin. Please ignore the birdhouse and don't touch it. |
| Environment class | CleanUpToysTask |
| Upstream name | CleanUpToysTask |
| Defined in | robolab/tasks/benchmark/clean_up_toys.py |
From https://github.com/NVlabs/RoboLab @ v0.3.1.
Tags¶
Task DomainManipulation
Skill primitives in the demo: semanticssorting
Why this task is interesting¶
Not yet written.
Capability notes¶
Not yet written.
Oracle demo review¶
Not yet reviewed.