featurebench/featurebench-modal
Coding
FeatureBench’s full task suite executed on Modal’s cloud sandbox runner.
Run this task
CLI:
inspect eval inspect_harbor/featurebench_modal --model openai/gpt-5Python:
from inspect_ai import eval
from inspect_harbor import featurebench_modal
eval(featurebench_modal(), model="openai/gpt-5")Dataset information
| Harbor registry | featurebench/featurebench-modal |
| Inspect task | featurebench_modal |
| Latest digest | sha256:a73b18b72e135c18de4477805e55fa3e87708f020bd5c875b64f86965f174e41 |
| Samples | 200 |
| Paper | arxiv |
See Task Parameters for the parameter set shared across all Harbor tasks.