Skip to content

Dataset · harborframework

Terminal-Bench 2.1 Benchmark for Harbor

harborframework/terminal-bench-2.1

Mirror of Terminal-Bench 2.1 benchmark tasks adapted for execution in Harbor.

What's inside

This is a private mirror of the harbor-framework/terminal-bench-2-1 repository containing tasks with text, images, and documents. The dataset includes under 1 thousand records and is prepared for direct download via git repository support in Harbor.

What you can do with it

  • Evaluating AI agents on terminal tasks
  • Automated model testing within the Harbor environment
Size
up to 1K rows
Downloads this month
260.4K
Likes
36

How to load it

from datasets import load_dataset

ds = load_dataset("harborframework/terminal-bench-2.1")