The authors define a generalist computer-use agent as a system in which the model decides what to observe, how to act and how to organize work across visual and programmatic interfaces.
Open agent stack improves computer tasks beyond GUI-only control
Across eight models, combining visual interaction, shell access and code execution improved OSWorld-v2 scores while reducing the number of reported steps.
Chinese Tech
Hanzhang Zhou · Long Li · Quyu Kong · Lei Wang · Panrong Tong · Long Chen · +9 more
Alibaba Token Hub · Alibaba Group
Research Digest··2 min read
Zhou et al.
§