Which AI Engines Cite Which Sources? ChatGPT vs Perplexity vs Gemini vs Claude

AI search is not one channel

ChatGPT, Perplexity, Gemini, and Google's AI features each draw from a different corpus and weight source types differently. A page one engine cites is often invisible to the others. In a BrightEdge analysis, the overlap between the sources two engines cite for the same topics ran as low as 16%. So "optimize for AI search" is the wrong frame. You are optimizing for four or five distinct retrieval systems at once.

This matters because the instinct, "rank on Google and the AI will follow," mostly does not hold. The engines disagree with Google, and they disagree with each other. If you want to understand the discipline itself, start with our AI visibility guide; this piece is the evidence layer underneath it: who cites what, and why it barely transfers.

What each engine actually pulls from

The clearest dataset comes from Profound's study of 680 million citations and Semrush's analysis of 100M+ citations. The headline: the same question produces structurally different sources depending on the engine.

ChatGPT leans encyclopedic and authoritative. Wikipedia is its single most-cited source at 47.9% of top sources, followed by Reddit and established media like Forbes. Historically its web results lean on Bing's index, so a page Bing has not indexed is hard to surface, and OpenAI's licensing deals with publishers like the Associated Press and the Financial Times shape what it can quote. Its mix is not static: Semrush watched ChatGPT's Reddit citations fall from roughly 60% of responses to about 10% over about a month in 2025, a shift that did not happen on the other engines.

Perplexity leans community and technical. Reddit is its top source at 46.7%, alongside expert and technical sources, and it runs its own crawler rather than relying solely on another engine's index.

Google's AI Overviews and AI Mode lean on UGC plus official sources. Pew Research found the most-cited sources in AI summaries are Wikipedia, YouTube, and Reddit, and that government sources appear in 6% of AI-summary sources versus 2% of standard results. LinkedIn shows up heavily for professional queries: Semrush measured it at 13.5% of Google AI Mode citations and 14.3% on ChatGPT, versus 5.3% on Perplexity.

Gemini is grounded in Google Search but behaves like its own surface, with low source overlap even against Google's own AI Mode.

Claude is the honest blank spot. It searches the web and returns citations, but credible, large-scale data on Claude's source-type distribution is not yet published. Rather than guess, treat it as a surface to monitor directly rather than model from secondhand numbers.

Ranking number one does not mean getting cited

This is where most SEO intuition breaks. Ahrefs studied 15,000 prompts and found that only 12% of the URLs cited by assistants like ChatGPT, Gemini, and Copilot rank in Google's top 10 for the original query. Per engine, the overlap with Google's top 10 ranged from about 28.6% for Perplexity (the outlier) down to 6 to 8% for ChatGPT. Semrush reached a similar conclusion from the other direction: the pages ChatGPT cites rank in organic positions 21 or lower almost 90% of the time, according to Semrush.

One important exception keeps this honest. Google's AI Overviews are not the same as the standalone assistants. Because they are built on Google's own results, they overlap with the ranking SERP far more, around 38% of AI Overview citations come from top-10 pages (down from roughly 76% in mid-2025), and BrightEdge tracked that overlap climbing from 32% to 54% over sixteen months. So Google rankings still strongly predict Google AI Overview citations, but they barely predict citations in ChatGPT, Perplexity, or Gemini.

The engines barely agree with each other

If the engines diverged from Google but agreed among themselves, you could optimize once. They do not. BrightEdge measured the pairwise overlap between engines' top cited sources at just 16% to 59%: Gemini and Google AI Mode shared only 27% of sources, while AI Mode and AI Overviews shared 59%.

There is a useful nuance inside that finding. Engines agree more on which brands to recommend than on which sources to cite. Brand-level overlap ran 36% to 55% even as source overlap stayed lower. The takeaway: building a recognizable, authoritative brand travels across engines, but a specific source-level win on one engine usually does not.

What this actually means for your strategy

Three moves follow directly from the data.

Build entity and brand authority everywhere, because that is the part that generalizes. Ahrefs' analysis of 75,000 brands found branded web mentions correlate with AI Overview visibility far more strongly (0.664) than backlinks do (0.218), and the top quartile of brands by web mentions earned roughly ten times the AI mentions of the next quartile. That is correlation, not a guarantee, but it points the same way as the brand-overlap data: presence and reputation across the web compound across engines.

Then optimize per engine for source-type fit. Authoritative, encyclopedic, well-structured content suits ChatGPT. Active, credible community and technical presence suits Perplexity. Classic ranking still matters most for Perplexity and Google's AI Overviews. Match the room you are trying to get into.

And track each engine separately, because source-level wins do not transfer. You cannot infer your Perplexity visibility from your ChatGPT visibility, or either from your Google rank. The practical workflow for that lives in our guide to monitoring your visibility across AI engines, and it is why cross-engine AI visibility tools exist (Frase among them). For the deeper question of why a page can rank yet stay uncited, see why your page ranks but isn't cited by AI, and for Google's own surfaces specifically, what Google's official AI guide says.

What to do now

Stop treating AI search as a single channel. Map where you actually appear engine by engine, double down on the brand and entity signals that carry across all of them, and tune each surface for the sources it favors.

Want to see who's cited where? Check your AI visibility across engines free and find the gaps before a competitor fills them.

Frequently asked questions

Do all AI engines cite the same sources?

No. Studies of hundreds of millions of citations show each engine draws from a different corpus, and the overlap between any two engines' cited sources runs roughly 16% to 59%. ChatGPT leans on Wikipedia and authoritative media, while Perplexity and Google's AI features lean more on community sources like Reddit. A source that wins on one engine often does not appear on another.

Does ranking number one on Google get me cited by AI?

Not reliably. Ahrefs found only 12% of URLs cited by assistants like ChatGPT, Gemini, and Copilot rank in Google's top 10 (Perplexity is the outlier at about 29%), and Semrush found ChatGPT cites pages ranking position 21 or lower almost 90% of the time. The exception is Google's own AI Overviews, which overlap with the ranking SERP much more (around 38%), so classic SEO predicts those citations far better than it predicts the standalone assistants.

Which sources does ChatGPT cite most?

Per Profound's data, Wikipedia is ChatGPT's most-cited source (about 48% of its top sources), followed by Reddit and established media such as Forbes. The mix shifts over time, so treat any single snapshot as a moving target rather than a fixed rule.

Which sources does Perplexity cite most?

Reddit is Perplexity's top source (about 47% in Profound's data), alongside technical and expert sources. Perplexity also has the highest overlap with Google's top-10 rankings among the standalone assistants (roughly 29%), so classic SEO tends to help here more than on ChatGPT.

How do I get cited by AI engines?

Build entity and brand authority across the web, since branded mentions correlate with AI visibility more strongly than backlinks, then tune each engine for the source types it favors and keep your content clearly structured and extractable. Because the engines weight sources differently, expect to optimize and measure per engine rather than once.

How do I know which AI engines are citing me?

You have to check each engine, because cross-engine visibility can't be inferred from one engine or from your Google rank. A cross-engine AI visibility checker shows where you appear across ChatGPT, Perplexity, Claude, Gemini, and Google's AI surfaces in one place.