Live Feed/Ahrefs/Fact Record
Ahrefs logo
Ahrefs
feature 96% Confidence Gate October 12, 2026

How ChatGPT Does Query Fan-Out (15 Million Queries Studied)

Ahrefs analyzed 66.6 million responses from the GPT-5.6 Luna model to investigate its query fan-out behavior. The study reveals that only 14.4% of responses trigger external search queries to retrieve information.

Verified State Diff

Comparison Mode:
- Previous State
The internal query fan-out logic and retrieval triggers of GPT-5.6 Luna were opaque to external observers.
+ Verified New State
Empirical data confirms that GPT-5.6 Luna utilizes a selective fan-out mechanism, triggering external queries in only 14.4% of logged-out user interactions.

Impact & Verification Analysis

WHO IS AFFECTED

SEO professionals, digital marketers, and content strategists.

WHY IT MATTERS

Understanding the specific trigger rate for AI search queries allows stakeholders to optimize content for AI visibility and predict how often their web properties are accessed by the GPT-5.6 Luna model.

Full Fact Overview

The analysis focuses on the internal retrieval mechanisms of the GPT-5.6 Luna model, specifically how it decides to perform 'fan-out' queries to external sources. By examining 15 million fan-out queries across 66.6 million interactions, the data provides empirical evidence on the frequency and triggers of AI-driven search behavior, which directly impacts SEO and content visibility strategies.

Multi-Source Evidence Chain (1)

How ChatGPT Does Query Fan-Out (15 Million Queries Studied)Ahrefs
TRACKED ENTITY
Explore all historical Ahrefs changes
View Ahrefs Hub ➔