ds-1000@head

Coding

DS-1000 is a code generation benchmark with 1000 realistic data science problems across seven popular Python libraries.

← Back to Registry

Run this task

CLI:

inspect eval inspect_harbor/ds_1000_head --model openai/gpt-5

Python:

from inspect_ai import eval
from inspect_harbor import ds_1000_head

eval(ds_1000_head(), model="openai/gpt-5")

Dataset information

Harbor registry ds-1000@head
Inspect task ds_1000_head
Version head
Samples 1000
Paper arxiv

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