Fruits On Plate¶
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 14% | 09-27 14:43 | 1h 30m | 162 | 16.0M / 95k | $0.231 | deterministic 1, grader_error 0, instance_exact 1, instance_ok 1, n_actions 138, video_rendered 1 |
| L✗ 1h 30m | 09-29 07:54 | 1h 30m | 416 | 60.1M / 177k | $0.761 | deterministic 1, grader_error 0, live_success 0, n_actions 3119, replay_success 0, video_rendered 1 |
Task instruction (upstream)
Put all the fruits on the plate
No oracle video published upstream — this is the scene it starts from.
What RoboLab states about this task
| Success predicate | object_on_top |
| Predicate arguments | object: ['lemon_01', 'lemon_02', 'lime01', 'lime01_01', 'orange_01', 'orange_02', 'pomegranate01'] · reference_object: clay_plates · logical: all · require_gripper_detached: True |
| Scored subtasks | 1 |
| Subtask predicates | pick_and_place_on_surface |
| Objects | lemon_01 lemon_02 lime01 lime01_01 orange_01 orange_02 pomegranate01 pumpkinlarge pumpkinsmall redonion serving_bowl clay_plates wooden_spoons spatula storage_box |
| Upstream attributes | semantics vague |
| Upstream difficulty | simple |
| Episode budget | 300 s |
| Other wordings | vague — Put fruits on the plate specific — Pick up all fruits that are on the table and place them all onto the large round plate |
| Environment class | FruitsOnPlateTask |
| Upstream name | FruitsOnPlateTask |
| Defined in | robolab/tasks/benchmark/fruits_to_plate.py |
From https://github.com/NVlabs/RoboLab @ v0.3.1.
Tags¶
Task DomainManipulation
Skill primitives in the demo: semanticsvague
Why this task is interesting¶
Not yet written.
Capability notes¶
Not yet written.
Oracle demo review¶
Not yet reviewed.