Overview

---

Introduction

PyRUA-Lean (Python for lean Robot-Use Agents, the robot counterpart of computer-use agents) turns a robot into a Python object. A tool-calling agent spends a full LLM call on every small step of a task and re-sends everything before it each time. A PyRUA-Lean agent writes code against robo instead, so one call can find an object, move above it, grasp, check the gripper and try again.

PyRUA-Lean has three parts so far:

• The framework. The agent composes robot primitives, VLA policies among them, into Python cells that check and retry locally and return only the images and state it asks for. On 700 task instances from LIBERO-PRO, RoboTwin 2.0 and RoboCasa365, with the same GPT-6 Astra planner and the same primitives as tool calling, success goes from 63.1% to 71.7%, with 49% fewer LLM calls and 65% fewer input tokens on the instances both agents solve.
• RoboDojo. With this library's own motion, gripper and perception primitives and the colour images of the benchmark's default observation, PyRUA-Lean reaches 51.62% average success rate (Score 58.20), first on all five dimensions of the public leaderboard (snapshot of October 9, 2026).
• PyRUA-Lean Coder v0.1-4B. Qwen3.5-4B trained with GRPO in the PyRUA-Lean framework, on 45 LIBERO-PRO instances, goes from 18.8% to 44.1% on 655 task instances it never saw in training, above the Qwen3.5-27B base model (29.2%).

Quick Start

git clone https://github.com/DAGroup-PKU/PyRUA-Lean.git && cd PyRUA-Lean
uv pip install -e .       # into the Python environment of your RPent robot stack (or pip)
pyrualean api             # the robot API the agent sees; no simulator needed

# one LIBERO-PRO episode: GPT-6 Astra writes code against robo, at most 40 code cells
export RPENT_ROOT=/path/to/rpent
pyrualean play --arm cells --suite libero_spatial_swap --task 7 --seed 2 \
    --model gpt-6-astra --max-turns 40 --budget-s 7200 --out runs/demo --cuda-device 0
pyrualean summarize runs/demo

LIBERO-PRO, RoboTwin 2.0 and RoboCasa365 run on RPent's robot stacks (the Quick Start sets one up step by step); RoboDojo runs through XPolicyLab and needs no RPent (RoboDojo).

Where to go next

• Quick Start

  Install, run one LIBERO-PRO episode, and look at what the agent did.

• How It Works

  What the agent writes, what robo returns, and what the model reads.

• Robots

  RoboDojo through XPolicyLab; LIBERO-PRO, RoboTwin 2.0 and RoboCasa365 on RPent. Each with the API reference its agent reads.

• PyRUA-Lean Coder v0.1-4B

  An open 4B model trained to write the cells, on Hugging Face.

Citation

If you use PyRUA-Lean, please cite our paper:

@article{si2026fewer,
  title={Fewer tokens, better action: GPT-6 Astra robot agents with 14\% higher success rate but 65\% fewer tokens},
  author={Si, Ruiyang and Bi, Jianxin and Yang, Shunyu and Ni, Rui and Huang, Wenbo and Wang, Qiang and Jiang, Shulong and Zhang, Yintianrun and Ruan, Feixiang and Zhang, Kewei and Wang, Duomin and Li, Xiuyu and Feng, Haiwen and Dong, Zhen and Zhou, Daquan},
  journal={arXiv preprint arXiv:2610.01939},
  year={2026}
}