OSWorld Verified

OSWorld Verified results
Provider	Model	Score	Source	Registry No.
Moonshot AI	Kimi K3 (max)	84.8%	Source for Kimi K3 (max) on OSWorld Verified (opens in a new tab)	120003
Moonshot AI	Kimi K2.6 (thinking)	73.1%	Source for Kimi K2.6 (thinking) on OSWorld Verified (opens in a new tab)	120002

Results