README.md

---
license: cc-by-4.0
pretty_name: LabDex
language:
  - en
tags:
  - robotics
  - embodied-ai
  - dexterous-manipulation
  - robot-learning
  - imitation-learning
  - laboratory-automation
---

# LabDex: A Hierarchical Benchmark for Dexterous Manipulation in Laboratories


<p align="center">
  <a href="https://zhipeng-tang.github.io/LabDex/">Project Page</a>
  &nbsp; | &nbsp;
  <a href="https://arxiv.org/pdf/2608.18618">Paper</a>
  &nbsp; | &nbsp;
  <a href="https://github.com/Zhipeng-Tang/LabDex">Code</a>
</p>

## Overview

**LabDex** is a large-scale dataset and benchmark for **dexterous manipulation in chemistry laboratories**, designed to support the training and systematic evaluation of robotic policies across different levels of task complexity.

LabDex provides unified **real-world and simulation platforms** and organizes laboratory manipulation into three hierarchical levels: **Atomic Skills**, **Compositional Skills**, and **Long-Horizon Laboratory Workflows**. The dataset contains **30+ atomic skills, 10+ compositional skills, 20+ laboratory assets, 8,000+ demonstration episodes, and over 1 million trajectory steps**.

Each demonstration includes synchronized **multi-view RGB observations, robot proprioceptive states, robot actions, and hierarchical task annotations**, providing a standardized foundation for research on laboratory dexterous manipulation and autonomous laboratory robots.

## Citation

```bibtex
@article{tang2026labdex,
  title={LabDex: A Hierarchical Benchmark for Dexterous Manipulation in Laboratories},
  author={Tang, Zhipeng and Chen, Sihang and Zhang, Sha and Yang, Peihao and Liu, Yan and Zhao, Wentao and Liu, Xinrui and Huang, Rui and Du, Wensheng and Huang, Yuting and Deng, Jiajun and Wang, Lidian and Zhang, Yuan and Zhang, Yanyong},
  year={2026},
  eprint={2608.18618},
  archivePrefix={arXiv},
  primaryClass={cs.RO}
}