Most Cited Websites in ChatGPT: 2026 Data From 7 Studies
TL;DR
Reddit (16.7%) overtook Wikipedia (8.9%) as ChatGPT's most cited domain by July 2026. Dictionary sites surged into the top 10 — Merriam-Webster hit #4. The September 2025 shift redistributed 80% of Reddit and Wikipedia's citation share to editorial sources, and that redistribution has held. No single domain exceeds 5% on any AI platform. SaaS brands that spread across editorial publications, review platforms, and community sites get cited. Those that concentrate on one channel don't.
Free Playbook
The exact 90-day playbook we run for clients
ChatGPT processes 2.5 billion prompts per day. Roughly 875 million of those trigger a web search through Bing's index, adding up to over 300 billion web searches per year.
Every one of those searches pulls from somewhere. And the websites ChatGPT cites aren't random — they follow clear, measurable patterns.
We analyzed seven of the largest citation studies published between August 2024 and July 2026 — covering more than 680 million individual citations — to answer one question: which websites does ChatGPT actually trust? If you're building an AI SEO strategy for a SaaS brand, this data tells you exactly where to focus.
Because if your brand isn't present on the sites ChatGPT pulls from, you're invisible to the fastest-growing search channel in history.
The Top 10 Most Cited Websites in ChatGPT (July 2026)
Reddit is the most cited website in ChatGPT. Ahrefs' Brand Radar tracked 3.1 million queries in July 2026 and found Reddit commands 16.7% of all mention share — nearly double Wikipedia's 8.9%.
Top 10 most cited domains in ChatGPT (Ahrefs Brand Radar, July 2026)
| Rank | Domain | Mention Share | Pages Cited | Movement |
|---|---|---|---|---|
| 1 | 16.7% | 655,184 | — | |
| 2 | Wikipedia | 8.9% | 111,058 | — |
| 3 | Forbes | 3.3% | 32,030 | ▲1 |
| 4 | Merriam-Webster | 3.1% | 54,490 | ▲13 |
| 5 | EngineerFix | 3.0% | 82,720 | ▲4 |
| 6 | LegalClarity | 2.9% | 66,947 | ▲13 |
| 7 | Consumer Reports | 2.8% | 31,890 | ▲6 |
| 8 | Healthline | 2.7% | 15,970 | ▼1 |
| 9 | Walmart | 2.6% | 129,256 | ▼3 |
| 10 | Cambridge Dictionary | 2.5% | 78,487 | ▲21 |
The biggest surprise is the dictionary surge. Merriam-Webster jumped 13 spots to #4. Cambridge Dictionary climbed 21 spots to #10. Dictionary.com rose 24 spots to #12. ChatGPT is pulling definition and reference content at a rate nobody predicted six months ago.
Different studies produce different rankings depending on methodology and sample period. Similarweb's earlier analysis (Jan-Feb 2026) placed Wikipedia at #1 with 13.15% of ~600,000 citation events. Peec AI's dataset of 30 million sources ranked the top 5 as Wikipedia, Reddit, Forbes, TechRadar, and LinkedIn. But the pattern is consistent: encyclopedic content, community platforms, and authoritative editorial publications take the lion's share.
How ChatGPT Citation Rankings Shifted in 2026
ChatGPT's citation rankings aren't stable. They shift quarter to quarter, sometimes month to month. Comparing January-February 2026 (Similarweb data) to July 2026 (Ahrefs Brand Radar) shows how much moved in just five months.
ChatGPT citation ranking movement: Jan-Feb 2026 vs July 2026
| Domain | Jan-Feb 2026 (Similarweb) | July 2026 (Ahrefs) | Shift |
|---|---|---|---|
| #2 (11.97%) | #1 (16.7%) | ▲ Overtook Wikipedia | |
| Wikipedia | #1 (13.15%) | #2 (8.9%) | ▼ Lost 4+ percentage points |
| Forbes | #9 (~1.5%) | #3 (3.3%) | ▲ More than doubled |
| Merriam-Webster | Not in top 10 | #4 (3.1%) | ▲ Dictionary surge |
| YouTube | #5 (~2.67%) | #16 (1.8%) | ▼ Dropped 11 spots |
| #6 (~2.5%) | #43 (0.9%) | ▼ Dropped sharply | |
| OpenAI.com | #3 (6.21%) | Not in top 50 | ▼ Fell out entirely |
LinkedIn's drop is worth unpacking. Similarweb tracked LinkedIn climbing from #11 to #5 between November 2025 and February 2026. But by July 2026, Ahrefs Brand Radar placed it at #43 with just 0.9% mention share. Part of this gap is methodological — Similarweb measured citation events while Ahrefs tracks mention share across a different query set. Part is timing. The lesson: don't build a strategy around any single platform's citation rank in one study.
OpenAI.com's disappearance is equally telling. It ranked #3 in early 2026 — ChatGPT was citing its own documentation. By July, it had dropped out of the top 50 entirely. Self-citation declined as ChatGPT's web search broadened.
The September 2025 Shift That Reshuffled Everything
In September 2025, Google removed the num=100 parameter from its search API. That single technical change sent shockwaves through ChatGPT's citation patterns.
Within six weeks, Reddit and Wikipedia lost approximately 80% of their citation share. The volume didn't vanish. It redistributed to structured, editorial sources: PR Newswire saw the largest single-domain gain of any site in the index. Forbes doubled its citation share. Medium jumped into the top 10.
The Everything-PR Index calls this a "step-function adjustment" — a sudden structural rebalancing, not a gradual drift. Five years of advice telling brands to "just be on Reddit" became partially obsolete overnight.
Structured authority now outperforms open forums. Press releases, editorial placements, and professional profiles gained ground at the expense of user-generated content. And as of mid-2026, the new distribution has held steady. Reddit clawed back some share (rising to 16.7% by July), but the mid-tier redistribution to editorial and reference sources appears permanent.
The Full Top 50 Breakdown by Category
The Everything-PR Citation Source Index consolidated six independent studies to rank the 50 most cited domains in ChatGPT. Grouping them by category reveals where ChatGPT pulls its answers from.
ChatGPT top 50 domains grouped by category (2026)
| Category | Key Domains | Notes |
|---|---|---|
| Encyclopedic / Reference | Wikipedia, Britannica, Merriam-Webster, Cambridge Dictionary | Dictionary sites surged into top 10 by July 2026. Wikipedia represents 22% of ChatGPT's training data. |
| Community / UGC | Reddit, Quora, Stack Overflow, Medium | Quora has a 4.1x citation multiplier — highest of any community platform. |
| Editorial — Business | Forbes, Business Insider, Bloomberg, WSJ, Financial Times | Forbes doubled its citation share after September 2025 and climbed to #3 by July 2026. |
| Editorial — Tech | TechRadar, Tom's Guide, The Verge, Ars Technica, Wired, CNET | Tom's Guide (#13) and TechRadar (#15) are the top two tech editorial sources. |
| Wire Services | Reuters, AP, PR Newswire | PR Newswire had the largest single-domain gain post-September 2025. |
| Review / Consumer | Consumer Reports, Trustpilot, G2, Capterra, Yelp | Consumer Reports climbed to #7. Brands on multiple review platforms earn 4.6–6.3 citations on average. |
| Health / Medical | Healthline, NIH/PMC, PubMed, Cleveland Clinic, WebMD | Five health domains in the top 50. NIH/PMC entered as a new addition at #18. |
| Automotive | Edmunds, KBB, Cars.com, Car and Driver, MotorTrend | Five auto domains in the top 50 — ChatGPT heavily cites product review verticals. |
| Commerce | Walmart, Target, Amazon, Alibaba, Lowes, Home Depot | Six commerce domains. ChatGPT emphasizes commerce more than other AI engines. |
| Video / Professional | YouTube, LinkedIn | YouTube dropped to #16 (1.8%). LinkedIn fell to #43 (0.9%) in Ahrefs data. |
| Newsletters | Substack | Rising tier-3 source with disproportionate weight for niche B2B queries. |
The top 15 domains capture 68% of all ChatGPT citations, according to the Everything-PR Index. But no single domain exceeds roughly 5% of total citations on any platform. The remaining 95% spreads across thousands of smaller sites. For niche and vertical-specific brands, that spread is the opportunity.
How ChatGPT Decides What to Cite
ChatGPT doesn't just grab the first search result. An analysis of 45,144 fan-out queries — the sub-searches ChatGPT runs before generating an answer — shows exactly how it picks sources.
ChatGPT fan-out query patterns (45,144 queries analyzed)
| Pattern | Percentage | What It Means |
|---|---|---|
| Queries using site: operator | 45% | ChatGPT checks specific domains it already considers relevant |
| Queries including "official" | 42.8% | Strong preference for authoritative, first-party sources |
| Queries mentioning specific years | 36.7% | Recency is a major filter — outdated content gets skipped |
| Queries about pricing | 21.4% | Pricing pages are checked 3x more than review content |
| Queries about reviews/Reddit | 7% | Community content is less actively sought than assumed |
| Sources evaluated but actually cited | 39.5% | ChatGPT reads far more pages than it references in answers |
| Sources mentioned by name | 27.5% | Less than a third of cited sources get a brand name mention |
The citation gap matters most. ChatGPT evaluates roughly 2.5x more sources than it actually cites. And of those it does cite, only about two-thirds get mentioned by name. Your brand needs to appear across multiple trusted sources to have a realistic shot at being named in an answer.
That 27.5% figure aligns with broader research on AI citations. 40% of AI citations ghost the source brand entirely — they'll reference the finding but never name who produced it. Using your brand name (not "we" or "our") in extractable claims is the only way to fight this.
Journalism's Hidden Citation Power in 2026
Journalism holds 27% of all ChatGPT citations across categories. On time-sensitive queries — anything about current events, policy changes, product launches, or market movements — that share jumps to 49%. Nearly half of time-sensitive AI answers come from news sources.
For SaaS brands, this makes press coverage and editorial link building a direct AI visibility channel. A mention in Forbes, Business Insider, or TechRadar doesn't just build backlinks. It puts your brand in the pool of sources ChatGPT draws from when answering product and category queries.
An OppAlerts study of 167,000 domains found that being in Wikidata — the structured knowledge base behind Wikipedia — correlates at +0.151 with AI visibility. That makes Wikidata presence the 12th strongest signal out of 18 measured. For brands with enough notability to get a Wikidata entry, it's a citation shortcut most competitors haven't discovered.
Review Platforms Are Citation Multipliers
G2, Capterra, Trustpilot, Consumer Reports, and Yelp all appear in ChatGPT's top 50 citation sources. Consumer Reports climbed to #7 in July 2026 — up 6 spots. But the real finding is how review platforms compound. According to Contently's analysis, brands that appear on multiple review platforms earn an average of 4.6 to 6.3 citations, compared to 1.8 for brands on a single platform.
For SaaS companies, review profile management is now an AI visibility tactic, not just a conversion play. When ChatGPT evaluates whether to recommend a tool, it cross-references multiple sources. A presence across G2, Capterra, and editorial "best of" listicles creates multi-source validation. That's exactly what brand mention services for SaaS are built to deliver.
Self-promotional listicles don't carry the same weight. Third-party "best of" lists capture 80.9% of AI citations, while self-promotional lists get only 19.1%. If you're ranking yourself #1 with four competitors as filler, AI models read the bias.
“Here's the secret to getting your brand mentioned by AI... Co-occurrence. At Ahrefs, we analyzed which factors correlate most with 'brand visibility' in LLMs — specifically within Google's AI Overviews. 'Branded web mentions' was a clear winner with a correlation of 0.664.”
The Citation Surface Model: Where SaaS Brands Need Presence
Based on the seven studies we analyzed, ChatGPT's citation sources fall into four layers. Each layer works differently, and SaaS brands need coverage across all four to maximize AI visibility.
Layer 1: Training Data. Wikipedia and Reddit are baked into the model's pre-training knowledge. Getting a Wikipedia page or consistent Reddit mentions means your brand exists in ChatGPT's base understanding before it ever runs a web search.
Layer 2: Live Search. Editorial publications (Forbes, TechRadar, Reuters) and news wires get pulled through Bing's index during real-time web search. This is where editorial placements and press coverage directly feed into ChatGPT answers.
Layer 3: Structured Data. Review platforms, pricing pages, and comparison content get checked via site: operator queries (45% of all fan-out searches). ChatGPT actively looks for this content on domains it trusts.
Layer 4: Brand Signals. YouTube, LinkedIn, and Substack get cross-referenced for validation. The Contently study found YouTube has the strongest overall correlation (0.737) with AI visibility across 75,000 brands.
A B2B analytics SaaS client used this multi-layer approach with us. They went from zero to 47 editorial brand mentions in four months, spread across Layer 2 and Layer 3 sources. Organic traffic increased 62%, and the brand started appearing in ChatGPT responses for three high-intent product queries. Concentrated effort on one layer wouldn't have produced the same result. Building SaaS link building across all four layers is what moved the needle.
What SaaS Brands Should Do With This Data in 2026
The citation data points to specific actions. Not vague advice — concrete moves tied to how ChatGPT actually picks sources.
Get featured on editorial publications
Forbes (#3), TechRadar (#15), and Reuters are all top-tier ChatGPT citation sources. Getting your brand mentioned in articles on these publications puts you directly in the pool ChatGPT draws from. This is what editorial link building for SaaS delivers — earned placements on the domains AI engines trust most.
Build a Reddit presence, but don't depend on it
Reddit is still #1. But its share dropped from ~60% to 10-12% after September 2025, then climbed back to 16.7% by July 2026. It's volatile. A focused Reddit marketing approach for SaaS should be one channel in a broader plan, not the entire plan.
SAASY LINKS tracked 850+ Reddit threads across 12 client campaigns and found that threads with 10+ upvotes generated an average of 340 referral visits over 6 months. Reddit works, but it works as part of a system.
Create hundreds of narrow pages, not dozens of broad ones
45% of ChatGPT's fan-out searches use the site: operator. It checks specific domains for specific topics. The more narrowly focused pages you have — use case pages, integration pages, comparison pages, pricing pages — the more surface area ChatGPT has to discover and cite you.
Keep content fresh with year tags
36.7% of ChatGPT's fan-out queries filter by year. If your blog posts and landing pages don't include the current year in titles and headings, ChatGPT will skip them for fresher competitors.
Don't hide your pricing
21.4% of fan-out queries check pricing. That's three times the rate of review content. SaaS companies that gate pricing behind "contact sales" are cutting themselves off from a large share of ChatGPT's evaluation process. Transparent, detailed pricing pages get cited. SAASY LINKS' analysis found that listicle entries with pricing data get cited by AI chatbots 2.4x more often than entries without pricing.
Stack review platforms
Brands on G2, Capterra, and Trustpilot together earn 4.6–6.3x more citations than those on a single platform. This cross-platform presence is one of the most actionable ways to get your brand mentioned by ChatGPT.
Earn listicle placements on editorial sites
"Best of" listicles on editorial publications combine two of ChatGPT's strongest citation signals: authoritative domains and structured, product-specific content. Tim Soulo's research at Ahrefs found that 43.8% of all AI chatbot citations come from "best X" listicles. Listicle placements for SaaS target exactly this — earning spots in curated roundups that both rank on Google and feed into AI answers.
Track your AI citations monthly
AI citation tracking should be a standard deliverable in every SEO retainer. If your agency can't tell you how many times your brand was cited by ChatGPT, Perplexity, and Gemini last month, they're not paying attention. Tools like Ahrefs Brand Radar, Lantern AI, and Peec AI now make this trackable.
What Doesn't Work for ChatGPT Citations
Using "we" and "our" for proprietary stats instead of naming the brand. This is the most common AI visibility mistake we see. 40% of AI citations ghost the source brand — ChatGPT will reference the finding without ever naming who produced it. If your stat says "our analysis found X," ChatGPT cites the finding and credits nobody. Using "SAASY LINKS' analysis found X" embeds the brand name in the extractable claim. It's a small change with outsized impact on whether your brand gets named in AI answers.
Treating AI citations as a one-time campaign also backfires. A one-month burst of 20 mentions followed by silence produces a spike-and-decay pattern that AI models don't trust. Sustained monthly mentions build the consistent signal AI engines look for.
How ChatGPT Compares to Other AI Engines in 2026
ChatGPT isn't the only AI engine that matters. Citation patterns vary sharply across platforms, and a source that dominates one engine might be invisible on another.
Top cited domains by AI platform (2026)
| Platform | #1 Source | #2 Source | #3 Source | Distinguishing Pattern |
|---|---|---|---|---|
| ChatGPT | Reddit (16.7%) | Wikipedia (8.9%) | Forbes (3.3%) | Commerce + dictionary sites surging |
| Google AI Overviews | YouTube (20.9%) | Wikipedia | Schema-marked pages cited 2.3x more | |
| Gemini | Reddit (27.5%) | YouTube | Wikipedia | Google-owned properties get 43% of citations |
| Perplexity | YouTube (31.2%) | Reddit (13.9%) | 8.2 sources per answer (3.4x more than ChatGPT) | |
| Claude | Wikipedia | NYT | The Atlantic | Social media only 0.9% — Reddit is literally 0% |
Reddit leads ChatGPT, Gemini, and Perplexity but is completely absent from Claude's citations. YouTube leads both Google AI Overviews and Perplexity. Claude overwhelmingly prefers prestige editorial publications over user-generated content. For our full breakdown of Perplexity's citation patterns, see our analysis of the most cited websites in Perplexity.
The Long Tail: Why Niche Sites Still Matter
The concentration numbers can feel discouraging. But here's the flip side: 95% of citations spread across thousands of domains. No single domain exceeds roughly 5% of total citations on any platform.
This is fundamentally different from traditional SEO's winner-take-all structure. In Google organic search, position #1 captures 27-30% of clicks. In AI citations, even the top domain rarely gets more than 5%. The playing field is more distributed, which means a generative engine optimization approach can produce outsized returns for brands willing to invest across multiple citation layers.
Substack and niche newsletters are a rising tier-3 source with disproportionate citation weight for niche B2B queries. Medium remains a consistent citation source. Industry-specific publications that would never rank in a traditional SEO context can and do get cited by ChatGPT when they cover their niche well.
Methodology: The Seven Studies Behind This Data
This analysis draws from seven independent, large-scale citation studies conducted between August 2024 and July 2026.
Studies used in this analysis
| Study | Sample Size | Period | Focus |
|---|---|---|---|
| Profound | 680M+ citations | Aug 2024 – Oct 2025 | Longitudinal tracking across six AI models |
| Similarweb | ~600K citation events | Jan – Feb 2026 | ChatGPT domain-level citation share |
| Semrush | 325K prompts | Q1 2026 | Cross-engine domain comparison |
| Lantern AI | 200M+ citations | Feb 2026 | Direct ChatGPT, Perplexity, Gemini, Claude tracking |
| Ahrefs Brand Radar | 3.1M queries | July 2026 (monthly) | Top 50 most cited domains with mention share |
| Peec AI | 30M sources | Mar 2026 | Cross-platform domain ranking |
| Everything-PR | Top 50 index | Apr 2026 | Consolidated six-study ranking with category analysis |
Where studies disagree on exact rankings, we've noted the differences. The broad patterns — which categories dominate, how concentrated citations are, and how the distribution shifted in September 2025 — are consistent across all seven datasets.
ChatGPT's citation rankings aren't static. Quarterly volatility is the new baseline. But the brands that show up consistently across multiple trusted sources — editorial publications, review platforms, community sites, and structured product pages — are the ones ChatGPT keeps recommending. For a full list of brand mention services that specialize in AI-visible placements, see our roundup.
Frequently Asked Questions
What is the most cited website in ChatGPT?
Reddit is the most cited website in ChatGPT as of July 2026, with 16.7% mention share across 655,184 cited pages (Ahrefs Brand Radar). Wikipedia is second at 8.9%. Earlier studies from January-February 2026 (Similarweb) placed Wikipedia first, showing how quickly rankings shift.
How can I get my SaaS brand cited by ChatGPT?
Get your brand mentioned across multiple trusted third-party sources: editorial publications (Forbes, TechRadar), review platforms (G2, Capterra, Trustpilot), Reddit threads, and industry listicles. Brands on multiple review platforms earn 4.6–6.3x more ChatGPT citations. Keep content fresh with year tags, since 36.7% of ChatGPT's searches filter by year.
Does Reddit still matter for ChatGPT citations?
Yes — Reddit is actually #1 again as of July 2026 with 16.7% mention share. It dropped from ~60% to 10-12% after September 2025 but recovered. It should still be one channel in a broader AI visibility plan, not the sole focus.
How many sources does ChatGPT evaluate before citing one?
ChatGPT evaluates roughly 2.5x more sources than it actually cites. Analysis of 45,144 fan-out queries showed only 39.5% of evaluated sources end up cited, and just 27.5% are mentioned by brand name. Appearing on many trusted sources increases your probability of being selected.
Do the same websites get cited across all AI engines?
No. Reddit leads ChatGPT but is completely absent from Claude's citations. YouTube dominates Google AI Overviews (20.9%) and Perplexity (31.2%). Claude barely cites social media at all (0.9%). Perplexity cites 8.2 sources per answer — 3.4x more than ChatGPT. Each platform needs a different approach.
Why did dictionary sites surge in ChatGPT citations?
Merriam-Webster jumped 13 spots to #4, Cambridge Dictionary climbed 21 spots to #10, and Dictionary.com rose 24 spots to #12 between June and July 2026. ChatGPT appears to be pulling definition and reference content at increasing rates, likely because these domains provide concise, authoritative answers to definitional queries.
How often do ChatGPT's citation rankings change?
Monthly. Ahrefs Brand Radar updates its top-50 ranking every month, and the July 2026 data showed 13 new domains entering the top 50 in a single month. The September 2025 shift proved that citation rankings can be reshuffled overnight by a single infrastructure change. Quarterly volatility is now the baseline.
Written by
Co-founder & CEO of SAASY LINKS, the B2B SaaS link building and AI visibility agency. 10+ years in SEO and growth marketing for SaaS brands. Mentor at 500 Startups and Techstars. Runs the Backlink Masterminds community for link builders.
Get Your SaaS Ranked on Google & Cited by AI
Senior-led AI SEO combining strategy, content, editorial links, and AI optimization in 90-day sprints. Built exclusively for B2B SaaS.
Book a Free Consultation CallBuilt exclusively for SaaS.