featurebench@1.0

Coding

FeatureBench full split: 200 feature-implementation tasks across 24 Python repos. 7 tasks require Ampere+ GPU.

← Back to Registry

Run this task

CLI:

inspect eval inspect_harbor/featurebench_1_0 --model openai/gpt-5

Python:

from inspect_ai import eval
from inspect_harbor import featurebench_1_0

eval(featurebench_1_0(), model="openai/gpt-5")

Dataset information

Harbor registry featurebench@1.0
Inspect task featurebench_1_0
Version 1.0
Samples 200
Paper arxiv

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