Skip to content

Sharpie marker · write c

keephardtabletop—0m 09s@JamesKrW

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 DexToolBench runs for every task of a run.

Codex + GPT-6 Luna, reasoning xhigh (ChatGPT login)default run

TrialStartedAgent timeModel requestsTokens in / outEst. costGrade
U✗ 1h 00m 0%09-28 23:361h 00m—12.0M / 165k$0.247grader_error 0, missing_trajectory 1
L✗ 1h 00m 0%09-29 08:161h 00m—22.6M / 173k$0.343deterministic 1, grader_error 0, video_rendered 1
1 other run(s) of DexToolBench

Codex + GPT-6 Luna, reasoning xhigh (Azure OpenAI API, 2026-09-30 stress test)

TrialStartedAgent timeModel requestsTokens in / outEst. costGrade
U✗ 59m 8%09-30 02:0759m—17.7M / 257k$0.348deterministic 1, grader_error 0, n_actions 2783, video_rendered 1
L✗ 59m 0% ⓘ10-01 02:5059m—14.2M / 129k$0.232deterministic 1, grader_error 0, video_rendered 1

Task instruction (upstream)

Pick up the sharpie marker and move it through the demonstrated motion (writing the letter C): bring it to each of the 25 goal poses in order.

Playback speed

Recorded by us: SimToolReal's pretrained RL policy replayed in our MuJoCo port of the task, four views (front, side / top, oblique); it reaches 25 of 25 goals here. The green ghost tool is the current goal.

What DexToolBench states about this task
Defined in dextoolbench/trajectories/marker/sharpie_marker/write_c.json
Category marker
Object sharpie_marker
Task write_c
N Goals 25
Tool Model assets/urdf/dextoolbench/marker/sharpie_marker/sharpie_marker.urdf
Table table_narrow_whiteboard.urdf (a whiteboard at the table's edge)
Metric task progress: goals reached / goals (8 grasp-box keypoints within 1.5 cm; 10 s per goal)

From https://github.com/tylerlum/simtoolreal @ 313d5ae.

Measured on the stack we run — MuJoCo 3.3.7 (CPU, bit-exact replay), KUKA iiwa 14 (MuJoCo Menagerie 1b86ece) + Sharpa HA4, 600 Hz physics, 60 Hz control
Harness Task dextoolbench_marker_sharpie_marker_write_c
Oracle SimToolReal's pretrained policy, recorded in this scene and replayed: 25/25 goals

Read from run in our MuJoCo port (robot_coding_bench tasks/dextoolbench_marker_sharpie_marker_write_c, image rcb-mujoco 0.1.1), 2026-09-28.

Tags

Task DomainManipulation

Why this task is interesting

Not yet written.

Capability notes

Not yet written.

Oracle demo review

Not yet reviewed.

Discussion