Mage
An open-source 4B-parameter desktop AI agent model from Microsoft that can autonomously decide whether to use graphical interface clicks or terminal commands to complete complex desktop tasks.
Timeline (last 90 days)
Microsoft has open-sourced Mage (called MajorVL Co-Work in the source material), a 4B-parameter desktop AI agent that scored 66% on the OSWorld Verified benchmark.
The human baseline on the OSWorld benchmark is about 72.4%, and Mage's 66% is close to that level, surpassing a model nearly 6 times its size (63.9%).
Mage's key capability is autonomously judging when to click the interface like a human and when to type terminal commands directly; it is 9.5GB in size and can run on a Mac mini.
All Facts (3)
The human baseline on the OSWorld benchmark is about 72.4%, and Mage's 66% is close to that level, surpassing a model nearly 6 times its size (63.9%).
50%UnverifiedMage's key capability is autonomously judging when to click the interface like a human and when to type terminal commands directly; it is 9.5GB in size and can run on a Mac mini.
50%UnverifiedMicrosoft has open-sourced Mage (called MajorVL Co-Work in the source material), a 4B-parameter desktop AI agent that scored 66% on the OSWorld Verified benchmark.
50%