Newest papers

Agent World Model in the Wild: Computer-Use Agent Environment Synthesis

Zhaoyang Wang, Qianhui Wu, Haorui Wang, Simon Yu, Baolin Peng, Canwen Xu, Hao Cheng, Wenlin Yao, Boyi Liu, Yite Wang, Jianfeng Gao, Huaxiu Yao · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.26828 · cited by 0 works (OpenAlex) · open-access PDF

Russian

Агент-обучение переходит от обучения от человеческих данных к обучению через взаимодействие с миром. Однако масштабирование такого обучения непосредственно в компьютерных системах реального мира является сложной задачей и может ввести существенные риски для безопасности. Поэтому мы предлагаем Агент World Model in the Wild (WildAWM), масштабируемый трубопровод для синтеза управляемых сред и задач для обучения общих агентов компьютерного использования (CUA). В то время как существующие подходы, как правило, специализируются на едином интерфейсе взаимодействия агентов, мы создаем исполняемые среды, которые совместно поддерживают графические пользовательские интерфейсы, услуги модельного контекста протокола и командные интерфейсы, позволяющие агентам гибко сочетать визуальное взаимодействие, использование инструментов и выполнение кода. WildAWM

Machine translation (Meta M2M100); the English original follows.

Abstract

Agent learning is shifting from learning from human data toward learning through interaction with the world. However, scaling such learning directly in real-world computer systems is challenging and can introduce substantial security risks. We therefore propose Agent World Model in the Wild (WildAWM), a scalable pipeline for synthesizing controllable environments and tasks for training general computer-use agents (CUAs). Whereas existing approaches typically specialize in a single agent interaction interface, we construct executable environments that jointly support graphical user interfaces, model context protocol services, and command-line interfaces, allowing agents to flexibly combine visual interaction, tool use, and code execution. WildAWM is entirely powered by open-source models, demonstrating that high-quality computer-use environments can be generated without proprietary frontier models. Beyond environments, we address three recurring challenges in task synthesis: ensuring solvability, maintaining sufficient difficulty, and producing reliable verifiers, by introducing execution-based quality gates that combine heterogeneous agent rollouts, cross-judge consistency checks, and counterfactual probes of judgment robustness. The resulting pipeline synthesizes 341 long-horizon environment-task pairs that can require over 100 agent steps to solve and remain challenging even for frontier models, with GPT-6-Astra solving only 60% of them. We train two Qwen models of different sizes exclusively in these synthetic worlds. Both models achieve substantial gains across 8 out-of-distribution benchmarks spanning web, claw, and computer-use agents, demonstrating that learning in WildAWM's synthesized worlds generalizes broadly to unseen computer-use tasks. The code is publicly available at https://github.com/MSR-Orchard/WildAWM.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Executable, Context (archaeology), Scalability, Pipeline (software), Artificial intelligence

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange