AndroidLab: Training and Systematic Benchmarking of Android Autonomous Agents

TL;DR

AndroidLab provides a comprehensive framework for training and benchmarking Android autonomous agents, supporting multimodal models and offering a standardized environment to improve success rates in Android tasks.

Contribution

It introduces AndroidLab, a systematic Android agent framework with a reproducible benchmark supporting LLMs and LMMs, and demonstrates significant performance improvements.

Findings

Success rates increased from 4.59% to 21.50% for LLMs.

Success rates increased from 1.93% to 13.28% for LMMs.

AndroidLab is open-sourced and publicly available.

Abstract

Autonomous agents have become increasingly important for interacting with the real world. Android agents, in particular, have been recently a frequently-mentioned interaction method. However, existing studies for training and evaluating Android agents lack systematic research on both open-source and closed-source models. In this work, we propose AndroidLab as a systematic Android agent framework. It includes an operation environment with different modalities, action space, and a reproducible benchmark. It supports both large language models (LLMs) and multimodal models (LMMs) in the same action space. AndroidLab benchmark includes predefined Android virtual devices and 138 tasks across nine apps built on these devices. By using the AndroidLab environment, we develop an Android Instruction dataset and train six open-source LLMs and LMMs, lifting the average success rates from 4.59% to 21.50% for LLMs and from 1.93% to 13.28% for LMMs. AndroidLab is open-sourced and publicly available at https://github.com/THUDM/Android-Lab.