featurebench/featurebench

Coding

FeatureBench: agentic coding on end-to-end feature-development tasks derived from open-source repositories.

← Back to Registry

Run this task

CLI:

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

Python:

from inspect_ai import eval
from inspect_harbor import featurebench

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

Dataset information

Harbor registry featurebench/featurebench
Inspect task featurebench
Latest digest sha256:313ee97ae5df80c618cf2818930ca140b3700b828cc336b602b9b19dbd9a59ff
Samples 200
Paper arxiv

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