Microsoft open-sources Mage-VL-Cowork, a 4B computer-use agent scoring 66.0% on OSWorld-Verified

Fri, Oct 30 20:13 UTC+8Updated Sat, Oct 31 18:32 UTC+8

Microsoft's Mage team released Mage-VL-Cowork weights on Hugging Face under Apache-2.0. The 4B-parameter vision-language model operates both the GUI (screenshots, mouse, keyboard) and code/terminal within a single policy. Microsoft says it scores 66.0% on OSWorld-Verified with a 50-step budget, above larger models such as Qwen3.6-27B; online RL adds 10.16 points over hybrid SFT.

Why it matters

A small, commercially usable open model approaching larger agents on computer use makes on-device agents more practical.

Sources2

Microsoft (Hugging Face)Official · WebsiteFri, Oct 30 20:13 UTC+8microsoft/Mage-VL-CoworkView original · via Hugging Face · Fri, Oct 30 20:13 UTC+8
MicrosoftMediaSat, Oct 31 18:32 UTC+8Mage-VL-Cowork: 4B Hybrid GUI + Code Computer-Use AgentView original · Sat, Oct 31 18:32 UTC+8

Written by dalao.ai: the title, summary and why it matters are our own words. The original reports are the record; their publishers keep authorship.