AI RESEARCH
OSExpert: Computer-Use Agents Learning Professional Skills via Exploration
arXiv CS.AI
•
ArXi:2603.07978v1 Announce Type: new General-purpose computer-use agents have shown impressive performance across diverse digital environments. However, our new benchmark, OSExpert-Eval, indicates they remain far less helpful than human experts. Although inference-time scaling enables adaptation, these agents complete complex tasks inefficiently with degraded performance, transfer poorly to unseen UIs, and struggle with fine-grained action sequences. To solve the problem, we