harbor-index/harbor-index-1.0

Harbor Index is an 82-task benchmark for agentic evaluation. It was distilled from more than 6,000 candidate tasks through repeated model runs, automated broken-task identification, human audits, and reward hacking supervision.

← Back to Registry

Run this task

CLI:

inspect eval inspect_harbor/harbor_index_1_0 --model openai/gpt-5

Python:

from inspect_ai import eval
from inspect_harbor import harbor_index_1_0

eval(harbor_index_1_0(), model="openai/gpt-5")

Dataset information

Harbor registry	harbor-index/harbor-index-1.0
Inspect task	harbor_index_1_0
Latest digest	sha256:9d4514cb93f6fafd9cf8ff352c784495ab675176c7f09671db523bd19b663584
Samples	82

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