Acrylic · Plumpark¶
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 RoboPaint 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 43m | 09-28 13:04 | 1h 43m | 141 | 19.1M / 127k | $0.279 | deterministic 1, grader_error 0, instance_exact 1, instance_ok 1, missing_trajectory 0, n_actions 70952, video_rendered 1 |
| L✗ 1h 22m 49% | 09-29 12:47 | 1h 22m | 139 | 19.3M / 158k | $0.300 | deterministic 1, grader_error 0, instance_exact 1, n_actions 71886, replay_success 0, video_rendered 1 |
Task instruction (upstream)
Copy the painting shown in the picture /app/target.png (a small acrylic painting after Utagawa Hiroshige's woodblock print Plum Park in Kameido, 1857) onto the main sheet, in at most 384 strokes.
Playback speed
Recorded by us: our reference solution, replayed by the verifier (canvas camera, scene camera, side camera, exact canvas).
What RoboPaint states about this task
| Success criteria | 1. scored at a 1 mm scale against the picture after the best single shift of the sheet, at most 5 mm along each axis 2. mean CIELAB ΔE <= 2.0 3. p90 ΔE <= 6.0 4. SSIM of L* >= 0.92 5. edge correlation >= 0.8 6. paint IoU >= 0.95 7. at most 384 strokes 8. both fresh-process replays of the trajectory meet this and end in the same state |
| Family | acrylic |
| Picture | a small acrylic painting after Utagawa Hiroshige's woodblock print Plum Park in Kameido, 1857 |
| Difficulty | hard |
| Tier | privileged |
| Continuous Score | 1 - mean ΔE / 20 |
| Grader Facts | name — plumpark alpha — truth_alpha.npz n_strokes — 295 source — painted by a hidden stroke program (solution/) with this instance's robot and brush, made by P's optimiser from File:De pruimenboomgaard te Kameido-Rijksmuseum RP-P-1956-743.jpeg license — CC0-1.0 area_mm2 — 50363.4 |
| Scoring | s.score() is served by a separate scoring container: it replays the actions the agent executed since its last reset on the graders' copy and returns the metrics, never images or target data |
| Agent Budget | 7200 s of wall clock per mode |
From our RoboPaint task definitions.
Tags¶
Task DomainManipulation
Why this task is interesting¶
Not yet written.
Capability notes¶
Not yet written.
Oracle demo review¶
Not yet reviewed.