Raise Platform¶
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 RoboWits 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 01m | 09-26 06:31 | 1h 01m | 107 | 12.0M / 111k | $0.199 | deterministic 1, grader_error 0, instance_exact 1, instance_ok 1, n_actions 1381, video_rendered 1 |
| L✗ 1h 30m | 09-29 21:59 | 1h 30m | 105 | 14.3M / 162k | $0.275 | deterministic 1, grader_error 0, live_success 0, n_actions 314, replay_success 0, video_rendered 1 |
Task instruction (upstream)
Place the book on the cup supports.
Playback speed
What RoboWits states about this task
| Success criteria | 1. Book is resting on both cup supports 2. Book is elevated above the table surface 3. Involved objects haven't fallen below the table |
| Objects | cup support 1 cup support 2 heavy book support cube 1 support cube 2 |
| Physics | rigid |
| Episode budget | 200 control steps |
| Default control mode | EE_ABS |
| Evaluation scenes | 50 |
| Published mutations | 4 |
| BlenderKit assets | 2 (not redistributable — mounted from the host) |
| Registry id | robowits/24-raise-platform-v0 |
| Environment class | RaisePlatformEnv |
| Defined in | gs_gym/envs/robowits/24_raise_platform.py |
From https://github.com/UMass-Embodied-AGI/RoboWits @ 9cc30ae.
Tags¶
Task DomainManipulation
Why this task is interesting¶
Not yet written.
Capability notes¶
Not yet written.
Oracle demo review¶
Not yet reviewed.