Scalable system verifies hybrid computer-use agents across platforms
RecreationWorld: Scalable and Verifiable Environments for Hybrid Computer-Use Agents
Computation and LanguageSoftware Engineering
Summary
Computer-use agents can either interact graphically or write code, but real digital tasks mix both. The authors created RecreationWorld, a system where agents learn by studying a working example on multiple platforms without fixed steps. The system tests how well agents can recreate software by watching and coding, rewarding them when their work matches the example. They also made a big set of tasks and tests, showing agents do better on interface design than on complex interactions or outputs. This helps move toward smarter software agents that can both use and build digital tools.
What this means in practice
- •For software developers: Test autonomous agents that combine interface use and coding to speed up workflow automation across multiple platforms.
- •For qa automation teams: Use a unified environment to validate software behavior by matching agent-produced outputs to verified references.
Authors
Shuai Bai, Jiayong Deng, Yikun Fu, Chang Gao, Xuhao Hu, Mianqiu Huang, Yizhen Jiang, Yuheng Jing, Dehui Kong, Keliang Li, Ning Li, Wanli Li, Dayiheng Liu, Dunjie Lu, Changwei Luo, Que Shen, Zheyuan Wang, Zijian Wang, Jie Wu, Gao Wu, Zhihui Xie, Rui Xie, Haiyang Xu, An Yang, Jiakang Yuan, Yanming Zhang, Jiajun Zhang, Xi Zhang, Zhenru Zhang, Zhuo Zhen, Mingkang Zhu, Bowen Zhou
Abstract
Computer-use agents (CUAs) have advanced along two separate lines: graphical interaction and software development through code and the command line. Real digital work requires both, interleaved rather than stacked end to end. We study hybrid CUAs that autonomously decide when to explore an interface, implement software, and run and visually verify their artifacts. We introduce RecreationWorld, a five-platform framework built around recreation: given a running reference, an agent must discover its behavior and build a faithful implementation with no prescribed workflow. RecreationWorld provides reproducible environments on Ubuntu, macOS, Windows, Android, and Web, plus a unified harness with native GUI control and coding tools. The running reference serves as an oracle for hidden behavioral tests, providing execution-grounded rewards. We scale trajectory generation with high-quality open-source applications. Models trained on these trajectories improve across five out-of-distribution coding and hybrid computer-use benchmarks and more frequently verify their rendered outputs, providing evidence of transfer beyond recreation. For held-out evaluation, we introduce RecreationBench, comprising 250 diverse tasks across domains and platforms. Reference-grounded programmatic and visual assertions cover action-conditioned outcomes at multiple interaction depths; each is validated on the reference and by human reviewers before the suite is frozen for automatic scoring. GPT-6 Astra leads at 58.1% overall, but passes all programmatic tests on just 2.8% of tasks. Agents reproduce static interface structure more reliably than interactions and computed outputs, while generated applications remain smaller and more monolithic than their references. We release the benchmark, environments, and test suites.