zenml/zenml-bench

Coding

ZenML Bench v0.1: 18 Harbor tasks where coding agents build, fix, configure and research with real ZenML 0.96.4; graded on the state ZenML recorded.

← Back to Registry

Run this task

CLI:

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

Python:

from inspect_ai import eval
from inspect_harbor import zenml_bench

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

Dataset information

Harbor registry zenml/zenml-bench
Inspect task zenml_bench
Latest digest sha256:5b0d64fed74870782a4c03763e8a4821dae3213c67bd470ae42d6fa854e73aa7
Samples 18
Source https://github.com/zenml-io/zenml-bench

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