Tools Picking All Hammers¶
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 0% | 09-26 17:53 | 1h 30m | 210 | 22.0M / 164k | $0.368 | deterministic 1, grader_error 0, instance_exact 1, instance_ok 1, n_actions 1032, video_rendered 1 |
| L✗ 1h 30m | 09-29 07:50 | 1h 30m | 160 | 19.7M / 120k | $0.285 | deterministic 1, grader_error 0, live_success 0, n_actions 1655, replay_success 0, video_rendered 1 |
Task instruction (upstream)
Take out all the hammers and put them on the table
No oracle video published upstream — this is the scene it starts from.
What RoboLab states about this task
| Success predicate | object_outside_of_and_on_surface |
| Predicate arguments | object: ['husky_hammer', 'blue_hammer', 'red_hammer', 'wood_hammer'] · container: left_bin · surface: table · logical: all · require_gripper_detached: True |
| Scored subtasks | 1 |
| Subtask predicates | pick_and_place_on_surface |
| Objects | clamp cordless_drill spring_clamp husky_hammer blue_hammer red_hammer wood_hammer left_bin center_bin right_bin clamp_01 |
| Upstream attributes | semantics spatial |
| Upstream difficulty | simple |
| Episode budget | 240 s |
| Other wordings | vague — Empty the hammers bin specific — Remove every hammer from the left bin and place each one on the table in front of you |
| Environment class | ToolsPickingAllHammersTask |
| Upstream name | ToolsPickingAllHammersTask |
| Defined in | robolab/tasks/benchmark/tools_picking_all_hammers.py |
From https://github.com/NVlabs/RoboLab @ v0.3.1.
Tags¶
Task DomainManipulation
Skill primitives in the demo: semanticsspatial
Why this task is interesting¶
Not yet written.
Capability notes¶
Not yet written.
Oracle demo review¶
Not yet reviewed.