Clean a Patio¶
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 BEHAVIOR-1K runs for every task of a run.
Codex + GPT-6 Luna, reasoning xhigh (ChatGPT login)default run
Not in this run: not among the 30 chosen for the first round (2026-09-26): mostly particles, cutting and liquids, or long tidy-ups like chosen ones
Task instruction (upstream)
Pick up the broom in the garden and sweep the mud off the patio floor until the floor is no longer covered in mud.
Measured on the stack we run — BEHAVIOR-1K v3.9.2 / OmniGibson 3.9.2 / Isaac Sim 5.1
| Goal predicates | covered |
| Goal clauses | 1 |
| Objects named in the goal | floor.n.01 mud.n.03 |
| Objects in the problem | 4 in 4 categories |
| Rooms loaded | corridor_0 garden_0 kitchen_0 living_room_0 |
| Demonstrations | 200 teleoperated episodes |
| Mean episode | 6m 42s (12,071 control steps at 30 Hz) |
| Base travel, mean | 37.21 m |
| Gripper travel, mean | left 11.73 m · right 11.25 m |
| Evaluation instances | 20 public test instances (ids 301–320) |
Read from BDDL activity definitions + 2026-challenge-task-instances (licensed download), 2026-09-21.
Tags¶
Why this task is interesting¶
Not yet written.
Capability notes¶
Not yet written.
Oracle demo review¶
Not yet reviewed.