Claw hammer · swing down¶
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
| Trial | Started | Agent time | Model requests | Tokens in / out | Est. cost | Grade |
|---|---|---|---|---|---|---|
| U✗ 54m 32% | 09-29 02:46 | 54m | — | 15.6M / 137k | $0.254 | deterministic 1, grader_error 0, n_actions 1386, video_rendered 1 |
| L✗ 1h 00m 0% | 09-29 07:09 | 1h 00m | — | 26.1M / 124k | $0.364 | deterministic 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)
| Trial | Started | Agent time | Model requests | Tokens in / out | Est. cost | Grade |
|---|---|---|---|---|---|---|
| U✗ 1h 00m 3% ⓘ | 09-30 20:46 | 1h 00m | — | 15.1M / 159k | $0.263 | deterministic 1, grader_error 0, n_actions 2246, video_rendered 1 |
| L✗ 59m 0% ⓘ | 09-30 14:35 | 59m | — | 19.8M / 126k | $0.280 | deterministic 1, grader_error 0, video_rendered 1 |
Task instruction (upstream)
Pick up the claw hammer and move it through the demonstrated motion (a downward hammer swing): bring it to each of the 37 goal poses in order.
Recorded by us: SimToolReal's pretrained RL policy replayed in our MuJoCo port of the task, four views (front, side / top, oblique); it reaches 37 of 37 goals here. The green ghost tool is the current goal.
What DexToolBench states about this task
| Defined in | dextoolbench/trajectories/hammer/claw_hammer/swing_down.json |
| Category | hammer |
| Object | claw_hammer |
| Task | swing_down |
| N Goals | 37 |
| Tool Model | assets/urdf/dextoolbench/hammer/claw_hammer/claw_hammer.urdf |
| Table | table_narrow_nail.urdf (one nail) |
| 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_hammer_claw_hammer_swing_down |
| Oracle | SimToolReal's pretrained policy, recorded in this scene and replayed: 37/37 goals |
Read from run in our MuJoCo port (robot_coding_bench tasks/dextoolbench_hammer_claw_hammer_swing_down, image rcb-mujoco 0.1.1), 2026-09-28.
Tags¶
Why this task is interesting¶
Not yet written.
Capability notes¶
Not yet written.
Oracle demo review¶
Not yet reviewed.