Pick Up Green Object¶
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 | 09-28 02:09 | 1h 30m | 128 | 14.9M / 141k | $0.246 | grader_error 0, missing_trajectory 1 |
| L✗ 1h 30m | 09-29 06:43 | 1h 30m | 213 | 26.7M / 205k | $0.447 | deterministic 1, grader_error 0, live_success 0, n_actions 1876, replay_success 0, video_rendered 1 |
Task instruction (upstream)
Pick up the green vegetable block
No oracle video published upstream — this is the scene it starts from.
What RoboLab states about this task
| Success predicate | object_picked_up |
| Predicate arguments | object: frozen_vegetable_block · surface: table |
| Scored subtasks | 1 |
| Subtask predicates | Subtask |
| Objects | frozen_vegetable_block blackandbrassbowl_large screwtoppail_a01 utilityjug_a02 |
| Upstream attributes | color |
| Upstream difficulty | simple |
| Episode budget | 30 s |
| Other wordings | vague — Pick up the green object specific — Identify the green vegetable block on the table and lift it off the table |
| Environment class | PickUpGreenObjectTask |
| Upstream name | PickUpGreenObjectTask |
| Defined in | robolab/tasks/benchmark/pick_up_green_object.py |
From https://github.com/NVlabs/RoboLab @ v0.3.1.
Tags¶
Task DomainManipulation
Skill primitives in the demo: color
Why this task is interesting¶
Not yet written.
Capability notes¶
Not yet written.
Oracle demo review¶
Not yet reviewed.