Mentions by LLM

This page tracks how often Wikipedia is mentioned across the 5 major AI chatbots: ChatGPT, Claude, Gemini, Grok and Perplexity. The data is based on the analysis of over 385'000 AI-generated responses and is updated monthly. Select a month below to see the detailed breakdown by model.

Key figures

2.69
citations per 100 AI answers
1st
of the 37 tracked platforms
ChatGPT
cites Wikipedia the most (6.99 per 100)
0.13
per 100 answers in September 2026

Wikipedia citations by AI model

ModelAnswers analyzedCitations of Wikipediaper 100 answers
ChatGPT114'4667'9966.99
Gemini114'4582'1481.88
Perplexity113'9621370.12
Claude22'54780.04
Grok22'0381210.55
All models387'47110'4102.69

Per 100 answers: citations of Wikipedia for every 100 answers of the model, from August 2025 to September 2026.

Wikipedia citations per 100 answers, each month, in French and in English.

MonthFrenchEnglish
Answersper 100 answersAnswersper 100 answers
August 20251'1405.79110*0.00
September 20252'7508.22375*5.87
October 20255'16010.79765*14.77
November 20252'73010.55255*17.65
December 20258'8407.36410*19.51
January 202622'4608.67240*10.42
February 202615'4209.91--
March 202626'6017.71--
April 202616'1830.12--
May 202619'8190.94--
June 202623'3362.26--
July 202655'3211.6610'5441.05
August 202664'1041.2013'3691.11
September 202678'8140.1016'2500.23

Hollow points and asterisk: fewer than 1'000 answers that month in that language, so the share varies strongly from one month to the next.

Frequently asked questions

How often do AI assistants cite Wikipedia?

On average, Wikipedia gets 2.69 citations per 100 answers from ChatGPT, Claude, Gemini, Grok and Perplexity, across 387'471 answers analyzed from August 2025 to September 2026. That puts Wikipedia 1st of the 37 platforms tracked by Sorank.

Which AI cites Wikipedia the most?

ChatGPT, with 6.99 citations per 100 answers. Claude cites Wikipedia the least: 0.04 per 100 answers.

Do AI assistants cite Wikipedia more in English or in French?

In English, Wikipedia gets 1.37 citations per 100 answers (42'318 answers). In French, 2.86 per 100 answers (342'678 answers).

Where do these figures come from?

From the questions Sorank asks AI assistants every week to track the visibility of its clients' websites: 387'471 answers since August 2025. Only aggregated figures are published; no client is named. Method

Wikipedia and AI Citations: The Encyclopedia That Powers Artificial Intelligence

No single source has shaped the knowledge embedded in modern AI systems more than Wikipedia. Since the earliest days of natural language processing research, Wikipedia's open, structured, and continuously updated articles have served as a foundational training resource for language models. Today, when users interact with AI chatbots like ChatGPT, Claude, or Perplexity, a substantial portion of the factual knowledge in their responses can be traced back to Wikipedia content.

Why Wikipedia Is the Top Source for AI Models

Wikipedia occupies a unique position in the information ecosystem. It is freely licensed, which means AI companies can legally use its content for training without copyright concerns. It is structured with consistent formatting (headings, infoboxes, categories and citation links), which makes it easy for models to parse and learn from. And it is massively multilingual, with editions in over 300 languages.

Data from Sorank shows that Wikipedia is consistently the single most-cited domain across all major AI chatbots. Perplexity explicitly links to Wikipedia articles with very high frequency. ChatGPT and Claude draw heavily on Wikipedia's knowledge during training.

What This Means for Businesses and Organizations

For businesses, the connection between Wikipedia and AI visibility creates both an opportunity and a challenge. Having a well-maintained Wikipedia article about your company significantly increases the likelihood that AI models will reference you accurately. The challenge is that Wikipedia has strict notability requirements and conflict-of-interest policies. Companies cannot simply create their own Wikipedia pages. Instead, the path runs through genuine notability: earning coverage from reliable independent sources.

The Wikipedia-to-AI Pipeline

When a company has a Wikipedia article, that article enters the training data for models like ChatGPT and Claude. It also appears in real-time search results that Perplexity queries. The information then shapes how the AI describes the company. Inaccurate or incomplete Wikipedia articles can lead to AI models generating misleading responses about a business.

Wikipedia is one of the platforms AI assistants cite most often. Track your citation data across ChatGPT, Claude, Perplexity, and other models using Sorank.

What is Generative Engine Optimization (GEO)?

GEO is the practice of optimizing your website so it gets mentioned and recommended by AI chatbots like ChatGPT, Claude, Gemini, Grok and Perplexity. Unlike traditional SEO which focuses on Google rankings, GEO focuses on how AI models reference your content when answering user questions. The data on this site tracks which domains are most mentioned across 385'000+ AI responses.

AI cites Wikipedia. Now it's your site's turn.

Sorank generates SEO-optimized articles and earns high-authority backlinks to rank your website on Google and in AI answers.

+5'000 users
Try for Free