Why ChatGPT and Perplexity Cite Completely Different Sources
ChatGPT and Perplexity agree on almost nothing when it comes to citations. Here's what the data actually shows, and what it means for where you focus your AEO work.
ChatGPT and Perplexity overlap on only an estimated 11% of cited domains, and just 12% of AI-cited URLs also rank in Google's top 10. They pull from genuinely different pools: ChatGPT leans heavily on Wikipedia, Perplexity leans on Reddit and very recent content. If you're optimizing for "AI search" as one target, you're likely under-optimizing for both.
Most AEO advice talks about "getting cited by AI" as if ChatGPT, Perplexity and Google AI Overviews all work the same way underneath. They don't, and the gap between them is bigger than most people assume.
The overlap is smaller than you'd think
Recent analysis of AI citation behavior found that only around 11% of domains get cited by both ChatGPT and Perplexity for the same kinds of queries. Zoom out further and only about 12% of URLs cited across AI tools also appear in Google's own top 10 organic results, meaning the large majority, roughly 88%, are pulling from sources that never rank on page one of a traditional search.
That second number is the one worth sitting with. A page can be genuinely well-optimized for classic SEO, ranking on page one, and still be functionally irrelevant to what an AI engine decides to cite, because the AI isn't just re-ranking the same candidate pool Google ranks. It's drawing from a different one.
Where each engine actually pulls from
| ChatGPT | Perplexity | |
|---|---|---|
| Average citations per answer | 10.4 | 21.9 |
| Top source preference | Wikipedia (47.9% of sources) | Reddit (46.7% of sources) |
| Freshness weighting | Less pronounced | Content under 30 days old gets 3.2x more citations |
| .edu share | Lower | Highest of the major engines (3.2%) |
| Country-code TLD share | Lower | Highest (4.4%) |
Perplexity isn't just citing more sources per answer, it's citing a meaningfully different kind of source: fresher, more community-driven, more international. ChatGPT leans toward the single most authoritative-seeming reference it can find, which in practice means Wikipedia shows up constantly.
Why they diverge this much
AI search isn't a single ranking step, it's a pipeline: the engine rewrites your query, retrieves a set of candidate passages, scores them, then synthesizes an answer with citations attached. Each engine tunes that pipeline differently, and those tuning choices are exactly what produces such different source pools. Perplexity's retrieval step is evidently weighted harder toward recency and community discussion; ChatGPT's leans toward encyclopedic authority.
This matters for anything you'd check with a free LLM Access Check: even if ChatGPT's and Perplexity's crawlers can both technically read your page fine, technical readability was never the thing separating them in the first place. The gap is in what each engine's retrieval step considers worth pulling from at all.
What this means for your own AEO work
A few concrete implications:
- Don't optimize for "AI search" as a single target. If Perplexity citations matter to you, freshness and community-adjacent credibility (real discussion, real recency) matter more than they do for ChatGPT. If ChatGPT citations matter more, being the clear, canonical, well-structured reference on a topic matters more.
- A page ranking well on Google tells you almost nothing about AI citation odds, given only 12% of AI-cited URLs overlap with Google's top 10. Check separately; don't assume.
- Freshness isn't a vanity metric for Perplexity specifically. A 3.2x citation multiplier for content under 30 days old is a real, measurable lever, not a soft "keep content updated" platitude.
If you want a page-specific read on this rather than the general pattern, a Full AEO Audit scores freshness, entity clarity and structural readability as separate pillars, exactly the dimensions this kind of citation gap tends to come down to. For the underlying distinction between optimizing for a ranked list versus optimizing for extraction and citation, see AEO vs SEO vs GEO.
Sources: 5WPR, State of AI Citations 2026; Attrifast, AI Search Ranking Factors 2026.
FAQ
If a page ranks well on Google, will it also get cited by ChatGPT or Perplexity?
Not reliably. Only about 12% of URLs cited by AI engines also appear in Google's top 10 organic results, so the two are far more independent than most people assume.
Should I optimize differently for ChatGPT versus Perplexity?
To some degree, yes. Perplexity weights recency and community/discussion sources (like Reddit) much more heavily, while ChatGPT leans on encyclopedic authority sources like Wikipedia. A single generic 'AI SEO' approach underserves both.
Does this mean technical crawler access doesn't matter?
It still matters, and it's the precondition: if a crawler can't read your page at all, none of this applies. But once a page is technically readable, the gap between engines comes from what each one's retrieval step chooses to pull, not from readability alone.