xlang-ai/osworld-verified

Assistants
Multimodal

OSWorld-Verified Harbor adaptation: 361 Ubuntu desktop-control tasks from OSWorld, excluding the 8 Google Drive/login credential-backed tasks.

← Back to Registry

Run this task

CLI:

inspect eval inspect_harbor/xlang_ai_osworld_verified --model openai/gpt-5

Python:

from inspect_ai import eval
from inspect_harbor import xlang_ai_osworld_verified

eval(xlang_ai_osworld_verified(), model="openai/gpt-5")

Dataset information

Harbor registry xlang-ai/osworld-verified
Inspect task xlang_ai_osworld_verified
Latest digest sha256:1402d7ce0c95876f51e3e532a4a4663b1d535db6a3afc22815e9272cef822a7f
Samples 361
Paper arxiv
Source https://github.com/xlang-ai/OSWorld

See Task Parameters for the parameter set shared across all Harbor tasks.