Top Sites Showing Up in ChatGPT (Across Multiple Studies)

• Reddit and Wikipedia dominate ChatGPT visibility across nearly all major datasets analyzed.
• Forbes, TechRadar, and Business Insider consistently appear, suggesting strong domain authority and topical relevance.
• OpenAI data partnerships don’t guarantee increased visibility in ChatGPT outputs.
• Query type and dataset design dramatically influence which sites are surfaced by AI.
• Sites with a strong presence on forums (e.g., Reddit) and review platforms (e.g., G2) may gain significant AI visibility.
• Traditional SEO fundamentals, such as authority, topical relevance, and reputation, should ultimately drive visibility in AI tools.

Several lists of the top sites that appear in ChatGPT are circulating at this point.

It’s getting so noisy at this point that I’m not really sure what these lists mean. After all, each study employs a different sample size and approach to the types of queries that comprise the study.

This is both good and bad. On the one hand, we get a more diverse group, and on the other, we don’t know who to trust.

That said, after reviewing these four reports, some sites appear to be universally visible.

In this post, I’ll show you the sites I’m seeing, what we can learn from these lists, and then dive deep into those that potentially overlap with OpenAI data partnerships.

Methodology:

I used studies from four different sources —Ahrefs, SEMRush, Profound, and Wellows —to compile a list of over a million queries.

To facilitate comparison, I added rankings where they were not initially present. I also added the percentage of total citations based on the provided dataset, if not offered.

1. Top 10 Sites Appearing on ChatGPT, according to Profound

Published June 2025

Profound is an AI visibility tool. Their database is one of, if not the largest. According to their study, their list of sites is based on a total of 680 million citations and displays the ranking as a percentage of total citations.

Here is what their top 10 looks like, broken down by percentage of citations.

What to learn from this?

They don’t get into their methodology here for the breakdown of the kinds of citations, but as you can see, Wikipedia is by far the largest source of citations.

Wikipedia accounts for almost half of the citations among the top 10 domains..

Then there is a steep drop-off to Reddit, Forbes, and G2.

Wikipedia is the largest source of free information on the web. It is also one of the primary sources of training data used in LLMs, such as GPT.

It’s also one of the hardest ones (at least I’ve found) to manipulate. However, it stands to reason that all brands should strive to establish a Wikipedia presence if possible.

Next, we’ll look at another extensive list from Ahrefs.

2. Top 10 Sites Appearing on ChatGPT, according to Ahrefs

Published September 2025

Ahrefs studied over 9.6 million queries and compiled the top 100 cited domains, making it the most extensive list available.

Although they list the top 100 in their dataset, I would like to focus on the top 10 so that we can compare the lists later. (Also, as you get further down the list, you end up with domains that are fractions of a percentage of the total makeup of the list.)

Note: I’ve added the percent of total (total mentions ≈2.88M).

What to learn from this?

As you can see, Reddit citations make up almost one-third of the list.

Then, Wikipedia also tops the list, followed by an incredibly steep drop-off to Amazon, and then many publisher sites.

You’ll notice a similar pattern in most of these studies (except for one), where Reddit dominates.

And this is definitely reflected in search results:

In March 2024, Reddit announced a partnership with OpenAI. The way they put it in their announcement:

“OpenAI will bring Reddit content to ChatGPT and new products, helping users discover and engage with Reddit communities. To do so, OpenAI will access Reddit’s Data API, which provides real-time, structured, and unique content from Reddit.”

Therefore, it certainly appears to be the case.

The following dataset, however, didn’t include any Reddit or Wikipedia data, so let’s explore that.

3. Top 10 Sites Appearing on ChatGPT, according to Wellows

Published July 2025

Next, based on the size of the study, is Wellows’ report, which analyzed 7,785 queries and 485,000 citations.

Wellows is essentially an AI agent that helps users build content.

Their post lists the top 50 sites, but their top 10 looks like this:

What to learn from this?

As you can see, there is a distinctly different look to this list: it’s missing Reddit and Wikipedia. The two sites don’t even appear on their top 50 list.

So, while Reddit may have the partnership, it seems like they certainly don’t show up everywhere.

Since this data comes from Wellow’s Kiva users (Kiva is the platform that helps generate content), their dataset may be biased as to the kinds of queries users ask, which is why you are seeing sites like Techradar, Cnet, and PCMag.com.

When I spoke with Masab Gadit, Founder of Wellows, he told me, “In our research, around 73% of the queries we analyzed were commercial in nature, and in that set, we didn’t see Reddit being cited much.

SEMrush, on the other hand, published different findings, but my assumption is that it’s influenced by the type of queries they used, their approach seems to be more from a niche/industry perspective, while ours was more content-type driven. So I see these as complementary rather than competing perspectives.”

That said, keep an eye on Techradar, because as you’ll see, it is also included on a few other lists.

4. Semrush AI Visibility Index

Published September 2025

Semrush has also released a study called the Semrush AI Visibility Index. It’s an incredibly well-designed analysis, though somewhat smaller in comparison to the other sites.

What I really like about their index, however, is that they analyzed what looks like a total of 2,500 prompts and responses across five major industries:

Note: Semrush reports these as index values (not percentages), so higher numbers = more visibility within that vertical.

As you can see, Reddit and Wikipedia top every list except for the “Digital Technology and Software” industry, where Wikipedia is third.

What to learn from this?

Once we get into the industry breakdowns, we encounter significant diversity in the other types of sites, aside from Reddit and Wikipedia, of course.

For instance, when searching for SaaS comparisons, users are most likely shown sites like G2, which is one of the most well-known software comparison and review platforms.

The Most Recurring Sites Cited by ChatGPT

Although all of these sites show up at a much different rate in every study, if we just look at the sites showing up on the list at all, the most commonly recurring sites on ChatGPT are:

Or here it is visualized:

To clarify, this is based on appearance across studies: Forbes (4), Reddit (3), Wikipedia (3), TechRadar (3), and Business Insider (3).

By visibility when present, Reddit and Wikipedia dominate.

(Note: Wellows’ dataset skews commercial/content-type and didn’t include Reddit/Wikipedia in its top 50, which explains the 3/4 count.)

What do all of these sites have in common?

They are all well-known sites with universal brand recognition.

Just by looking at the overlap here, we can get a pretty good sense of how and where ChatGPT tends to fill the gaps when Reddit and Wikipedia aren’t enough.

However, as you may or may not know, OpenAI has data partnerships with many publisher sites. In the next section, I aim to investigate whether these partnerships lead to increased exposure.

Do Sites with OpenAI Data Partnerships Result in Greater Exposure in ChatGPT?

While some sites are suing OpenAI over their use of their content, many publishers have struck data partnerships with OpenAI. Here’s a look at the list of all of the data partnerships as of the time of publication:

1. Associated Press
2. Axel Springer
3. Le Monde
4. Prisa Media
5. Financial Times
6. News Corp
7. Reddit
8. Dotdash Meredith
9. Vox Media
10. The Atlantic
11. Condé Nast
12. Future plc
13. Hearst
14. The Guardian
15. Axios
16. The Washington Post
17. Time USA

Or, here it is visualized:

In a previous study, I compared this list with data from a Ziff Davis study that looked into sites used to curate ChatGPT datasets. There, the overlap was minimal.

For this study, I wanted to determine whether publisher sites (or sites within the media companies) under these partnerships received any preferential treatment from OpenAI.

First, I had to break down the sites that exist under these media companies (for instance, Future PLC has 50+ sites under its umbrella).

Then, looking at the top 10 in every list, I put together this matrix to show which study cited each brand and the rank in which they appeared:

(Since SEMrush had broken it down by industry, I averaged the results to obtain the raw counts of the rank.)

If it’s marked N/A, there are no sites in the top 10.

Here it is in visual format:

When you roll this up to the publisher group level, not the site level, here is what it looks like:

Note: Averages are based only on appearances, not on absence.

Alternatively, if you prefer a visual breakdown…

Ultimately, there was less of a connection than I thought there would be.

Out of the 17 confirmed partners, only 8 have any Top 10 visibility in the studies.

And of those 8, many came from the SEMRush study, which provided more industry-specific insights.

And the visibility that exists is middling.

What Can We Learn From All of This?

Ultimately, these lists are starting to tell the story that many SEOs and marketers need to know. It’s not about partnerships or hacking the algorithm.

Reddit and Wikipedia are Almost Universally Visible

Clearly, Wikipedia and Reddit dominate across almost every dataset. And afterwards, there is a steep drop-off in visibility.

For example, in Profound, Wikipedia accounts for nearly 8% of all citations, while Reddit accounts for ~1.8%, and everything else is a fraction of a percent.

This means a small handful of sites command a disproportionate share of visibility.

The Type of Query Changes the Picture

Given that Wellows has no Reddit or Wikipedia entries in its dataset, we can see that query set design matters significantly: the kinds of prompts you ask, who is asking them, and how.

In the SEMrush industry breakdown set, the industry-specific breakdowns revealed new sites that we hadn’t seen in other lists, such as G2 in SaaS and Investopedia.

Forbes is the Most Recurring Site (behind Reddit and Wikipedia)

It doesn’t account for a massive percentage in the lists it appears on, but Forbes is actually the most commonly recurring site across all the lists (behind Reddit and Wikipedia).

Then, Techradar.com and Businessinsider.com. (Nothing against Techradar, but that one surprised me the most!)

So are these the ones to pitch if you want to show up in citations?

I wouldn’t think that way.

So What Actually Matters for Visibility in ChatGPT?

It appears that all the same factors we consider as SEOs also matter for visibility in ChatGPT, as well as a few new ones we may want to incorporate into our purview.

The existing skillset:

• Building topical relevance (we saw topically-relevant sites show up in SEMRush’s industry-specific study).

• Naturally raising domain authority (we saw Forbes, Reuters, and other high authority sites showing up across the board),

Newer skillset:

• Managing online reputation on forums and review sites (clearly, Reddit plays a big part in this, as do sites like G2).

• Then, when we add Wikipedia, I think that citations from there speak to the need for brand awareness and ubiquity.

These don’t sound like rocket science, but I think because we have such limited data, we are all grasping at whatever strategies we can and trying to scale them.

However, what I’ve outlined are the fundamentals of marketing.

To show up in ChatGPT and survive in the AI era, we simply need to be good marketers; it’s that straightforward.

Vince Nero

Vince is the Director of Content Marketing at Buzzstream. He thinks content marketers should solve for users, not just Google. He also loves finding creative content online. His previous work includes content marketing agency Siege Media for six years, Homebuyer.com, and The Grit Group. Outside of work, you can catch Vince running, playing with his 2 kids, enjoying some video games, or watching Phillies baseball.

More Posts by Vince Nero
Website: https://buzzstream.com