OSWorld2.0：面向长程现实任务的计算机操作代理基准测试

摘要

现有计算机使用基准无法捕捉真实计算机使用的现实性、复杂性和长时域需求，限制了其揭示前沿智能体局限性的能力。我们推出OSWorld 2.0基准，包含108个涵盖日常与专业任务的长时间计算机使用工作流，旨在捕捉复杂且具挑战性的真实世界现象。每个任务代表一个真实端到端工作流，人类用户完成其中位时间约1.6小时，而使用最大思考能力的Claude Opus 4.7平均需调用318次工具（OSWorld 1.0中约30次）。OSWorld 2.0聚焦于真实工作流中常见但此前基准未充分体现的挑战性现象，涵盖交互设计挑战（如流式交互与动态环境）以及智能体模式挑战（如跨源推理、隐状态推断与视觉空间精度）。任务基于真实输入工件，并与有状态用户画像数据进行交叉参照，同时包含独立的安全报告以审计安全敏感的执行过程。在我们500步的主二值完成度指标下，采用最大思考与批量工具调用的Claude Opus 4.8表现最佳，但仅完成20.6%的任务（部分得分54.8%）；GPT-5.5令牌效率远高于前者，但得分停滞在13%附近。这些结果表明，当前智能体距离专业级计算机使用仍相去甚远：它们并非在基本图形界面控制或编码上出错，而是会丢失约束条件、忽略任务中途出现的信息、猜测而非询问用户、跳过验证，在需要恢复隐藏状态的关键任务上表现最为挣扎。

English

Existing computer-use benchmarks fail to capture the realism, complexity, and long-horizon demands of real-world computer use, limiting their ability to reveal the limitations of frontier agents. We introduce OSWorld 2.0, a benchmark of 108 long-horizon computer-use workflows across everyday and professional tasks, designed to capture complex and challenging real-world phenomena. Each task represents a realistic end-to-end workflow that takes human users a median of about 1.6 hours to complete and requires an average of 318 tool calls with Claude Opus 4.7 using maximum thinking, compared with about 30 in OSWorld 1.0. OSWorld 2.0 targets challenge phenomena that are common in real workflows yet underrepresented in prior benchmarks, spanning interaction-design challenges such as streaming interaction and dynamic environments, as well as agent-pattern challenges such as cross-source reasoning, implicit-state inference, and visual-spatial precision. Tasks are grounded in authentic input artifacts and cross-referenced against realistic stateful user profile data, and include separate safety reports auditing safety-sensitive execution. Under our primary binary-completion metric at 500 steps, Claude Opus 4.8 with maximum thinking and batched tool calls scores best but still completes only 20.6% of tasks at a 54.8% partial score; GPT-5.5 is far more token-efficient yet plateaus near 13%. These results show that current agents are still far from professional-level computer use: rather than stumbling on basic GUI control or coding, they lose track of constraints, miss information that arrives mid-task, guess rather than ask the user, and skip verification, struggling most when a task hinges on hidden state they must recover.