About this paper
A simplified evaluation harness that keeps full episode history and stops motions early substantially improves robot-use agent scores on RoboDojo, with scene geometry providing the largest additional benefit.
GPT-6 Astra can drive a robot arm directly through a fixed harness, and scores of such robot-use agents on RoboDojo have risen quickly since. Reported systems differ in harness and test subset and often change several designs at once, so it is hard to tell which choices produce the gains. We take these designs apart and test them in one framework, along the harness around the model, the action interface, the reuse of experience and the scene geometry given to the model. A lean harness that keeps the full episode history and stops each motion early goes a long way. Reusing experience changes how the model acts but is not a free lunch, and scene geometry helps most, even when depth is…
GPT-6 Astrarobot-use agentsevaluation harnessL3 harnessJustRUAaction interfaceexperience reusescene geometrydepth prediction from RGBlearned policy
- Upvotes
- 3
- Authors
- 12
- Published
- Nov 11, 2026
Summary and abstract excerpt via Hugging Face Papers ↗ · arXiv.
Trending history
JustRUA: A Controlled Study of GPT-6 Robot-Use Agents was first seen on Hugging Face · Daily papers on Nov 13, 2026. Since then it has appeared in 28 checks on 4 different days. Its best position was #26 of 30, reached on Nov 14, 07:28 UTC. It is #29 right now, climbing from #30 to #29 over the last 6 checks.