ModelRefs / Needle in a Haystack — AI Glossary

Needle in a Haystack — AI Glossary

A long-context benchmark inserting a specific fact ('needle') at varying positions in a long document and measuring retrieval accuracy. Also called NIAH.

Overview

NIAH tests whether a model can retrieve a single injected fact from a distracting document corpus at any context position. Frontier models (Gemini 1.5 Pro, Claude 3 Opus) approach near-perfect NIAH scores at 128K context. Failures reveal positional biases in attention patterns. Extended to multi-needle and multi-hop variants.

Reference details

Topicprompting
Also known asNIAH
Last reviewed2026-06-24

Continue your research

Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to Needle in a Haystack — AI Glossary.

Frequently asked questions

What is Needle in a Haystack?

A long-context benchmark inserting a specific fact ('needle') at varying positions in a long document and measuring retrieval accuracy.

Is Needle in a Haystack the same as NIAH?

Yes — NIAH are common aliases for Needle in a Haystack.

What concepts are related to Needle in a Haystack?

Closely related concepts include lost in the middle, long context, retrieval pipeline.