[KongchangAI]
BenchmarkOS World 2.1 / OS World

OS World 2.1

A benchmark environment for evaluating a model's ability to operate graphical user interfaces (GUIs); its tasks simulate a person completing actions on a desktop system such as clicking menus, filling out forms, and switching between applications.

Timeline (last 90 days)

Nov 1

GPT-6.1 Terra comprehensively surpasses the previous generation on the visual reasoning benchmark MMMU Pro, and on the desktop operation benchmark OS World 2.1 it reaches the previous generation's top-tier score at medium reasoning effort.

Unverified50%

All Facts (1)

Unverified

GPT-6.1 Terra comprehensively surpasses the previous generation on the visual reasoning benchmark MMMU Pro, and on the desktop operation benchmark OS World 2.1 it reaches the previous generation's top-tier score at medium reasoning effort.

50%

Source Articles

GPT-6.1 Terra Launches at Rock-Bottom Price: Do Flagship Work for a Tenth of the CostNov 1GPT-6.1 Terra Hands-On: Blazing Fast, but Outperformed by DeepSeek?Nov 1