Saved in:
Bibliographic Details
Main Authors: Qin, Zengyi, Chen, Jinyuan, Man, Yunze, Cao, Shengcao, Pang, Ziqi, Wang, Zhuoyuan, Fang, Han, Zhu, Ling, Xie, Zixin, Wei, Zibu, Ran, Tianshu, Geng, Haoran, Pan, Ray, Sun, Qizhen, Bright, Zachary, Cai, Yuyang, Yang, Chongye, Zhao, Jiace, Liu, Tianrui, Cao, Han, Zhou, Yeyang, Wang, Rui, Wang, Song, Ren, Xiang, Zhang, Bo, Ban, Yutong, Abbeel, Pieter, Anthony, Brian
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2511.11672
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910095732899840
author Qin, Zengyi
Chen, Jinyuan
Man, Yunze
Cao, Shengcao
Pang, Ziqi
Wang, Zhuoyuan
Fang, Han
Zhu, Ling
Xie, Zixin
Wei, Zibu
Ran, Tianshu
Geng, Haoran
Pan, Ray
Sun, Qizhen
Bright, Zachary
Cai, Yuyang
Yang, Chongye
Zhao, Jiace
Liu, Tianrui
Cao, Han
Zhou, Yeyang
Wang, Rui
Wang, Song
Ren, Xiang
Zhang, Bo
Ban, Yutong
Abbeel, Pieter
Anthony, Brian
author_facet Qin, Zengyi
Chen, Jinyuan
Man, Yunze
Cao, Shengcao
Pang, Ziqi
Wang, Zhuoyuan
Fang, Han
Zhu, Ling
Xie, Zixin
Wei, Zibu
Ran, Tianshu
Geng, Haoran
Pan, Ray
Sun, Qizhen
Bright, Zachary
Cai, Yuyang
Yang, Chongye
Zhao, Jiace
Liu, Tianrui
Cao, Han
Zhou, Yeyang
Wang, Rui
Wang, Song
Ren, Xiang
Zhang, Bo
Ban, Yutong
Abbeel, Pieter
Anthony, Brian
contents Training computer use agents requires full-featured OS sandboxes with GUI environments, which consume substantial hardware resources as the number of sandboxes scales. Stochastic errors arising from diverse software execution within these sandboxes further demand robust infrastructure design and reliable error recovery. We present OSGym, a scalable OS environment infrastructure for computer use agents, built around these key optimization strategies: (1) Decentralized OS state management, which isolates failures to individual replicas and significantly enhances overall system reliability; (2) Hardware-aware OS replica orchestration, which addresses CPU-bounded scaling bottlenecks and substantially reduces compute overhead; (3) KVM virtualization with copy-on-write disk management, which shares a common bootable disk across VM instances and provisions only instance-specific modifications, reducing physical disk consumption by 88% and increasing disk provisioning speed by 37 times; and (4) Robust container pool with multi-layer fault recovery. Together, these optimizations yield strong scalability and resource efficiency: OSGym manages over a thousand OS replicas under constrained resources, supports parallel trajectory generation at 1420 multi-turn trajectories per minute, and reduces per-replica cost to 0.2-0.3 USD per day, a 90% reduction over standard deployment. Our experiments validate OSGym across end-to-end pipelines for data collection and training for computer use agents. We believe OSGym establishes a new foundation for scalable, general-purpose computer use agent research.
format Preprint
id arxiv_https___arxiv_org_abs_2511_11672
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle OSGym: Scalable OS Infra for Computer Use Agents
Qin, Zengyi
Chen, Jinyuan
Man, Yunze
Cao, Shengcao
Pang, Ziqi
Wang, Zhuoyuan
Fang, Han
Zhu, Ling
Xie, Zixin
Wei, Zibu
Ran, Tianshu
Geng, Haoran
Pan, Ray
Sun, Qizhen
Bright, Zachary
Cai, Yuyang
Yang, Chongye
Zhao, Jiace
Liu, Tianrui
Cao, Han
Zhou, Yeyang
Wang, Rui
Wang, Song
Ren, Xiang
Zhang, Bo
Ban, Yutong
Abbeel, Pieter
Anthony, Brian
Distributed, Parallel, and Cluster Computing
Training computer use agents requires full-featured OS sandboxes with GUI environments, which consume substantial hardware resources as the number of sandboxes scales. Stochastic errors arising from diverse software execution within these sandboxes further demand robust infrastructure design and reliable error recovery. We present OSGym, a scalable OS environment infrastructure for computer use agents, built around these key optimization strategies: (1) Decentralized OS state management, which isolates failures to individual replicas and significantly enhances overall system reliability; (2) Hardware-aware OS replica orchestration, which addresses CPU-bounded scaling bottlenecks and substantially reduces compute overhead; (3) KVM virtualization with copy-on-write disk management, which shares a common bootable disk across VM instances and provisions only instance-specific modifications, reducing physical disk consumption by 88% and increasing disk provisioning speed by 37 times; and (4) Robust container pool with multi-layer fault recovery. Together, these optimizations yield strong scalability and resource efficiency: OSGym manages over a thousand OS replicas under constrained resources, supports parallel trajectory generation at 1420 multi-turn trajectories per minute, and reduces per-replica cost to 0.2-0.3 USD per day, a 90% reduction over standard deployment. Our experiments validate OSGym across end-to-end pipelines for data collection and training for computer use agents. We believe OSGym establishes a new foundation for scalable, general-purpose computer use agent research.
title OSGym: Scalable OS Infra for Computer Use Agents
topic Distributed, Parallel, and Cluster Computing
url https://arxiv.org/abs/2511.11672