simpleqa@1.0

Knowledge

SimpleQA: 4,326 short, fact-seeking questions from OpenAI for evaluating language model factuality. Uses LLM-as-a-judge grading.

← Back to Registry

Run this task

CLI:

inspect eval inspect_harbor/simpleqa_1_0 --model openai/gpt-5

Python:

from inspect_ai import eval
from inspect_harbor import simpleqa_1_0

eval(simpleqa_1_0(), model="openai/gpt-5")

Dataset information

Harbor registry simpleqa@1.0
Inspect task simpleqa_1_0
Version 1.0
Samples 4326
Paper arxiv

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