What capabilities does AA-Omniscience test?

AA-Omniscience evaluates factual recall, hallucination.

What license is AA-Omniscience under?

AA-Omniscience is available under Closed.

AA-Omniscience

Active

Artificial Analysis's broad-knowledge benchmark - thousands of curated factual questions spanning specialized domains - designed to test hallucination calibration.

Open

Publisher: Artificial Analysis
Capabilities: Factual Recall Hallucination
Format: Custom
License: Closed
Published: Nov 2025
Notable for: Benchmark for evaluating factual recall and hallucination.
Canonical: artificialanalysis.ai/methodology/intelligence-benchmarking

Cite

Notes

Only stored in your browser.

Papers

AA-Omniscience: A Benchmark for Long-Tail Factual Knowledge

preprint · 2025

Artificial Analysis's benchmark of 6,000 expert-written long-tail factual questions across 42 economically relevant topics, scored with a hallucination-penalized metric.

introduces

AA-Omniscience: A Benchmark for Long-Tail Factual Knowledge

preprint · 2025

Artificial Analysis's benchmark of 6,000 expert-written long-tail factual questions across 42 economically relevant topics, scored with a hallucination-penalized metric.

FAQ

What is AA-Omniscience?: Artificial Analysis's broad-knowledge benchmark - thousands of curated factual questions spanning specialized domains - designed to test hallucination calibration.
What capabilities does AA-Omniscience test?: AA-Omniscience evaluates factual recall, hallucination.
What license is AA-Omniscience under?: AA-Omniscience is available under Closed.