Parse
IndexPricing
Sign in
  1. Research
Research
Perplexity vs ChatGPT Search vs Google AI Mode: 3-engine citation anatomy

Perplexity vs ChatGPT Search vs Google AI Mode: 3-engine citation anatomy

Published Nov 20, 2026Updated Nov 21, 2026

On this page
  • Overlap vs winners
  • 3-way overlap
  • Citation density
  • Homepages vs deep links
  • Editorial vs social diet
  • Full source diet
  • Top domains per engine
  • What to do
  • Study scope
  • How we measured
  • More like this
  • Overlap vs winners
  • 3-way overlap
  • Citation density
  • Homepages vs deep links
  • Editorial vs social diet
  • Full source diet
  • Top domains per engine
  • What to do
  • Study scope
  • How we measured
  • More like this

Dimitry ApollonskyFounder, Parse

DataWhat reviewed means

When a buyer asks the exact same question on Perplexity, ChatGPT Search, and Google AI Mode, do the three engines cite the same web pages, prefer the same page formats, and agree on who wins #1?

We matched 857 organic buyer questions answered by all three engines from November 1 through November 19, 2026 (2,571 answers and 19,727 total citation links), measuring pairwise URL and domain overlap, top-brand agreement, source-type mix, and URL path depth, alongside a September 1 to October 16 Claude (Opus 5) reference cohort (6,780 answers, 54,708 links).

Across 857 matched organic buyer questions answered by Perplexity, ChatGPT Search, and Google AI Mode in November 2026, all three engines named roughly seven brands per answer (7.19 to 7.46) yet shared a cited URL on only 3.85% of questions. Perplexity cited homepages only 1.77% of the time (15.0x lower than ChatGPT's 26.59%) and pulled editorial and news domains at 2.4x ChatGPT's share, while Google AI Mode cited social platforms 13.5x as often as ChatGPT.

Engines agree 3x to 4x more on the #1 brand than on the URLs behind it

Across 857 matched organic buyer questions answered by Perplexity, ChatGPT Search, and Google AI Mode in November 2026, pairwise URL overlap stays in single digits. Averaged across all 857 questions, mean URL Jaccard is 2.54% between Perplexity and ChatGPT Search, 1.87% between Perplexity and Google AI Mode, and 1.68% between ChatGPT and Google AI Mode (all-857 root-domain Jaccard: 8.39%, 4.77%, and 5.40%). Even restricted strictly to questions where a pair shares at least one URL, conditional URL Jaccard is only 7.58% (Perplexity vs ChatGPT), 9.81% (Perplexity vs Google), and 9.57% (ChatGPT vs Google), with conditional root-domain Jaccard at 13.34%, 13.72%, and 15.92%.

Yet the three engines agree on who wins #1 roughly one-third of the time. Where both engines assign a #1 direct brand rank, Perplexity and ChatGPT agree on the top brand 30.78% of the time (n = 783), ChatGPT and Google AI Mode agree 28.94% (n = 812), and Perplexity and Google AI Mode agree 26.85% (n = 797).

Perplexity vs ChatGPT overlap-only URL Jaccard
7.58%

9.81% Perplexity vs Google · 9.57% ChatGPT vs Google

All-857 means: 2.54% / 1.87% / 1.68%; all 3 share >=1 URL on 3.85%

Perplexity vs ChatGPT #1 brand agreement
30.78%

26.85% Perplexity vs Google · 28.94% ChatGPT vs Google

Engines converge far more on brand picks than URLs

Perplexity average answer length (9.97 cited URLs)
2,294 chars

vs 7,739 chars (10.82 URLs) on ChatGPT Search

Google AI Mode writes 3,444 chars with 2.23 URLs

Homepage share in Perplexity citations
1.77%

15.0x lower than ChatGPT Search (26.59%)

Google AI Mode sits between them at 6.17%

Conditional URL/domain overlap and #1 brand agreement across 857 questions

  • Perplexity vs ChatGPT
  • Perplexity vs Google AI Mode
  • ChatGPT vs Google AI Mode
  • URL Jaccard (overlap-only)
  • Perplexity vs ChatGPT7.58%2.54% all-857
  • Perplexity vs Google AI Mode9.81%1.87% all-857
  • ChatGPT vs Google AI Mode9.57%1.68% all-857
  • Domain Jaccard (overlap-only)
  • Perplexity vs ChatGPT13.34%8.39% all-857
  • Perplexity vs Google AI Mode13.72%4.77% all-857
  • ChatGPT vs Google AI Mode15.92%5.40% all-857
  • #1 brand agreement
  • Perplexity vs ChatGPT30.78%
  • Perplexity vs Google AI Mode26.85%
  • ChatGPT vs Google AI Mode28.94%
  • 0%10%20%30%40%
Nov 1 to Nov 19, 2026 · Conditional Jaccard overlap (%) over pairs sharing >=1 URL/domain (all-857 URL means: 2.54%, 1.87%, 1.68%) and #1 brand agreement (%).

Takeaway

Market consensus on brand winners travels across engines even when each engine cites completely different web pages to justify the pick.

All three engines cite the same exact URL on only 3.85% of questions

Looking for an exact page cited simultaneously by ChatGPT Search, Google AI Mode, and Perplexity turns up a hit on only 3.85% of the 857 matched questions. Broadening from exact URLs to root domains raises three-way overlap to 14.59%.

More than 8 in 10 questions produce completely disjoint domain lists across the three engines, which makes single-engine citation tracking blind to most cross-engine discovery.

3.85%
of matched questions share even one URL across all three engines
14.59% share at least one root domain (857 matched questions)

Takeaway

Winning a citation in ChatGPT Search does not get you cited in Perplexity or Google AI Mode; 96% of questions share zero URLs across all three.

Perplexity packs 9.97 citations into 2,294-char answers; Google cites 2.23 URLs

All three engines recommend nearly identical shortlist sizes: 7.42 brands per answer on ChatGPT Search, 7.19 on Perplexity, and 7.46 on Google AI Mode. Their citation budgets diverge completely.

ChatGPT Search writes long 7,739-character answers with 10.82 cited URLs (median 10) across 7.80 domains, or 1.46 citations per named brand. Perplexity compresses almost the same citation load (9.97 URLs, median 10, across 7.89 domains; 1.39 per brand) into tight 2,294-character answers with a 0.00% zero-citation rate. Google AI Mode writes 3,444-character answers with just 2.23 URLs (median 2; 0.30 citations per brand) and leaves 10.50% of answers with zero prose citations.

Answer length, brands named, and citation footprint across 857 matched questions

ChatGPT Search10.82107.87.421.461.057,739
Perplexity9.97107.897.191.3902,294
Google AI Mode2.2322.17.460.310.53,444
Nov 1 to Nov 19, 2026 · Latest answer per engine on 857 organic questions (2,571 answers).

Takeaway

Google AI Mode names seven brands on two citations, while ChatGPT and Perplexity back seven brands with roughly ten citations each.

ChatGPT cites homepages 26.59% of the time, 15.0x more often than Perplexity

The three engines point to completely different layers of a website. On ChatGPT Search, 26.59% of all 9,272 citation links land on a bare root homepage (with 9.39% on product/pricing paths and 5.33% on docs/help paths).

Perplexity almost never cites homepages: only 1.77% of its 8,544 links hit a root URL (15.0x lower than ChatGPT Search), and instead 25.60% point to /blog/, /news/, /article/, /insights/, or /resources/ paths (vs 8.09% on ChatGPT). Google AI Mode sits between them at 6.17% homepages and 22.03% blog/editorial paths.

URL path profile across 19,727 citation links on 857 matched questions

  • ChatGPT Search
  • Google AI Mode
  • Perplexity
  • Homepage URL (/)
  • ChatGPT Search26.59%
  • Google AI Mode6.17%
  • Perplexity1.77%
  • Blog or editorial path
  • ChatGPT Search8.09%
  • Google AI Mode22.03%
  • Perplexity25.60%
  • Product or pricing path
  • ChatGPT Search9.39%
  • Google AI Mode4.13%
  • Perplexity3.75%
  • Docs or help path
  • ChatGPT Search5.33%
  • Google AI Mode1.57%
  • Perplexity1.59%
  • 0%10%20%30%
ChatGPT Search (9,272 links), Google AI Mode (1,911 links), Perplexity (8,544 links).

Takeaway

Homepages carry a quarter of ChatGPT Search citations, while Perplexity and Google AI Mode overwhelmingly demand deep article and blog URLs.

Perplexity doubles down on editorial and news (18.78%); Google leans social (11.46%)

Each engine favors a distinct publisher tier. On Perplexity, editorial sites take 10.90% of citations (931 links) and news domains take 7.88% (673 links). Combined, editorial and news represent 18.78% of Perplexity's citations, 2.4x ChatGPT Search's combined 7.89% (5.28% editorial and 2.61% news).

Google AI Mode diverges toward user-generated social platforms: social domains account for 11.46% of Google AI Mode links (219 of 1,911), 13.5x ChatGPT Search's 0.85% (79 of 9,272) and 7.6x Perplexity's 1.51% (129 of 8,544). ChatGPT Search leads on documentation (6.84% vs 2.56% on Google and 1.18% on Perplexity) and government sources (6.99% vs 2.41% and 2.26%).

Distinguishing source categories across the three matched engines

  • ChatGPT Search
  • Google AI Mode
  • Perplexity
  • Editorial
  • ChatGPT Search5.28%
  • Google AI Mode6.96%
  • Perplexity10.90%
  • News
  • ChatGPT Search2.61%
  • Google AI Mode6.23%
  • Perplexity7.88%
  • Social
  • ChatGPT Search0.85%
  • Google AI Mode11.46%
  • Perplexity1.51%
  • Documentation
  • ChatGPT Search6.84%
  • Google AI Mode2.56%
  • Perplexity1.18%
  • 0%5%10%15%
Share of each engine's citation links on the 857 matched questions (Nov 1 to Nov 19, 2026).

Takeaway

Perplexity rewards editorial journalism, Google AI Mode pulls in social proof from Reddit and YouTube, and ChatGPT Search favors vendor docs and government sources.

Explore every source type across ChatGPT, Google, Perplexity, and Claude

Across all 19 classified source categories, general web and SaaS websites form the largest base on every platform, accounting for 53.64% combined on ChatGPT Search (29.86% general, 23.78% SaaS), 55.50% on Perplexity (32.54% general, 22.96% SaaS), and 48.51% on Google AI Mode (23.13% general, 25.38% SaaS).

For reference, Parse's separate September 1 to October 16, 2026 Claude (Opus 5) reference cohort (6,780 answers and 54,708 domain links) skewed even heavier toward general (41.30%) and SaaS (34.86%) domains, with 7.74% editorial and 3.52% news, and almost zero documentation (0.99%) or social (0.28%).

Citation share by source type across four engines

Select an engine to inspect its full source-type breakdown by link count and percentage share.

Share (%) · Top 10 of 19

  • General32.49%
  • SaaS22.93%
  • Editorial10.9%
  • News7.91%
  • Directory5.91%
  • Ecommerce4.11%
  • Reference3.84%
  • Government2.26%
  • Blog2.11%
  • Finance1.86%
  • 0%10%20%30%40%
Explore the data (19 rows)
PerplexityGeneral2,77632.49
PerplexitySaaS1,95922.93
PerplexityEditorial93110.9
PerplexityNews6767.91
PerplexityDirectory5055.91
PerplexityEcommerce3514.11
PerplexityReference3283.84
PerplexityGovernment1932.26
PerplexityBlog1802.11
PerplexityFinance1591.86
PerplexitySocial1291.51
PerplexityDocumentation1011.18
PerplexityDeveloper820.96
PerplexityHealth650.76
PerplexityEducation420.49
PerplexityCommunity240.28
PerplexityAcademic220.26
PerplexityEncyclopedia200.23
PerplexityVideo10.01
Matched 857-question cohort (Nov 1 to Nov 19) for ChatGPT Search, Google AI Mode, and Perplexity; separate Sep 1 to Oct 16 reference cohort for Claude (Opus 5).

Takeaway

Each engine keeps its own source signature, from Claude and ChatGPT leaning into SaaS and vendor sites to Perplexity favoring editorial publishers and Google pulling social.

Only Amazon and Microsoft make every engine's top 10 on matched prompts

The top 10 root domains on the exact same 857 questions confirm how little the three engines overlap. Only two domains (microsoft.com and amazon.com) appear in all three top-10 lists. ChatGPT Search's top 10 is filled with tech platforms, SaaS vendors, and regulators led by microsoft.com (255 links), google.com (212), amazon.com (189), splunk.com (85), cloudflare.com (83), and ftc.gov (58).

Google AI Mode leads with reddit.com (129 links) and youtube.com (44 links), followed by amazon.com (23), linkedin.com (21), and gartner.com (20). Perplexity leads with ones.com (194), parse.gl (110, reflecting AI-visibility prompts in the matched organic cohort), amazon.com (97), topconsumerreviews.com (93), and microsoft.com (82).

Top 10 root domains per engine on 857 matched questions

Switch engines to see the 10 most-cited root domains on the identical 857-question sample.

Links · 10 results

  • microsoft.com255
  • google.com212
  • amazon.com189
  • splunk.com85
  • cloudflare.com83
  • clickup.com74
  • elastic.co70
  • atlassian.com63
  • ftc.gov58
  • asana.com55
  • 0100200300
Explore the data (10 rows)
1ChatGPT Searchmicrosoft.com255
2ChatGPT Searchgoogle.com212
3ChatGPT Searchamazon.com189
4ChatGPT Searchsplunk.com85
5ChatGPT Searchcloudflare.com83
6ChatGPT Searchclickup.com74
7ChatGPT Searchelastic.co70
8ChatGPT Searchatlassian.com63
9ChatGPT Searchftc.gov58
10ChatGPT Searchasana.com55
Nov 1 to Nov 19, 2026 · Root-domain link counts across the 857 matched organic buyer questions.

Takeaway

Tailor your source strategy by engine rather than chasing one generic leaderboard.

What to do with these numbers

Track and build for each engine separately. With 2.54% all-question URL overlap (7.58% on overlapping prompts) between Perplexity and ChatGPT Search and 3.85% three-way URL overlap, winning citations on one platform rarely protects you on the others.

Match the page format each engine rewards: keep crisp homepages, product/pricing pages, and documentation hubs for ChatGPT Search (26.59% homepages, 6.84% docs); invest in deep blog, research, and editorial placements for Perplexity (25.60% blog/editorial paths, 18.78% editorial and news domains); and maintain credible YouTube and Reddit visibility for Google AI Mode (11.46% social share).

  • Compare competitor visibility across engines
  • Query cross-engine citations in your workspace via Parse MCP

Study scope and counting rules

The matched cohort includes all 857 organic buyer questions that received answers on all three engines (Perplexity, ChatGPT Search, and Google AI Mode) between November 1 and November 19, 2026. Using the latest recorded answer per engine per question yields 2,571 answers and 19,727 total citation links (9,272 on ChatGPT Search, 8,544 on Perplexity, and 1,911 on Google AI Mode). Because the matched organic prompts include AI visibility categories, Parse's own domain (parse.gl) appeared #2 in Perplexity's cohort (110 links) and was retained without filtering.

Pairwise URL and root-domain Jaccard overlap is computed per question (intersection divided by union of distinct cited URLs or root domains) and reported both across all 857 matched questions (2.54%, 1.87%, and 1.68% URL; 8.39%, 4.77%, and 5.40% root domain) and conditionally over questions where the two engines share at least one URL or root domain (7.58%, 9.81%, and 9.57% URL; 13.34%, 13.72%, and 15.92% root domain). Three-way overlap counts the share of questions where at least one URL (3.85%) or root domain (14.59%) appeared in all three engines' answers. #1 brand agreement counts matched rank-1 direct brand mentions (lowest mention position) where both engines assigned a #1 brand. URL path types classify normalized citation URLs by root path ('', '/') vs /blog|news|article|insights|resources/, /docs|documentation|help|support|kb|guide|learn/, and /product|products|pricing|features|solutions|platform/. The Claude (Opus 5) source-type profile (6,780 answers, 54,708 links from Sep 1 to Oct 16, 2026) uses a separate prompt cohort and window and is reported strictly as a reference baseline.

How we measured

We analyzed 857 matched organic buyer questions answered by ChatGPT Search, Google AI Mode, and Perplexity from November 1 through November 19, 2026 (2,571 answers; 19,727 total citation links), plus a 6,780-answer Claude (Opus 5) reference cohort (54,708 domain links, September 1 through October 16, 2026).

URL overlap between Perplexity and ChatGPT Search
7.58% (2.54% all)

7.58% on overlapping prompts (2.54% across all 857); all 3 share a URL on 3.85%

#1 brand agreement between Perplexity and ChatGPT Search
30.78%

3x to 4x higher than conditional URL Jaccard (7.58% to 9.81%)

homepage citation share: Perplexity vs ChatGPT Search
1.77% vs 26.59%

Perplexity favors deep blog/editorial URLs (25.60% vs 8.09%)

social domain citation share: Google AI Mode vs ChatGPT
11.46% vs 0.85%

Google AI Mode cites social platforms 13.5x as often

Get the data

Full study dataset (CSV)3-engine answer metrics, overlap rates, full source-type breakdown across 4 engines, URL path shares, and top 10 domains per engine.

More like this

Do ChatGPT and Google cite the same page?
Even when both engines cited the same website for the same question, they usually chose different pages.
ChatGPT Search vs Google AI Mode: which sources each engine cites
Only 1 of the top 10 root domains (amazon.com) overlaps between ChatGPT Search and Google AI Mode across 3.07M citations (Aug to Nov 2026).
ChatGPT Search vs Google AI Mode (2026): citation, brand, and source report card
Across 547K strict public answers (Aug-Nov 2026), ChatGPT Search cites 2.6x more pages per answer, flags reluctance 2.7x more often, and avoids social domains.
Does AI cite brand homepages?
67.6% of citations to a brand's own website pointed below the homepage, led by Google AI Mode at 92.1%.
AI citation Pareto curve across 4.18M links
0.39% of root domains (676 of 175,035) capture 50% of ChatGPT Search citations and 60.85% of cited URLs appear once, yet on matched prompts Google AI Mode concentrates its Top 100 tighter (42.00% vs 35.28% and 29.88%).
Google AI Mode Shopping Cards vs prose citations
97.8% of 57,765 Google AI Mode Shopping Card links route inside Google's closed overlay, 53.08% of card answers have zero prose citations, and Walmart and Target outnumber Amazon more than 7-to-1.
How AI recommends MCP and agent-native tools across 911,413 answers
94.96% of 6,292 MCP mentions across 911,413 answers happened unprompted (surging +127.12% on ChatGPT Search), while narrow agent/API topics flip #1 55.00% of the time.
PR and publisher tier ROI in AI search
On competitive multi-brand pages, wire press releases convert just 5.08% into #1 AI picks vs 17.29% for creators and 15.42% for affiliates.
For agents: this page as Markdown, https://parse.gl/research/perplexity-vs-chatgpt-vs-google-ai-mode-citations.md. JSON, MCP · free for the last 90 days, checks from $0.01
Open in ChatGPT (opens in a new tab) Open in Claude (opens in a new tab)

This page

As Markdown: https://parse.gl/research/perplexity-vs-chatgpt-vs-google-ai-mode-citations.md
As CSV: https://parse.gl/research/perplexity-vs-chatgpt-vs-google-ai-mode-citations/dataset.csv
Reuse: Figures and tables on this page are free to republish with attribution: a link to https://parse.gl/research/perplexity-vs-chatgpt-vs-google-ai-mode-citations beside them. A standard followed link (dofollow) is preferred.

Read and search

Any brand's shares and rivals, with 95% ranges: GET https://agent.parse.gl/v1/tools/check_brand?subject=%3Cbrand-domain%3E
Third-party cited pages missing a brand: GET https://agent.parse.gl/v1/tools/search_cited_pages?args=%7B%22pages%22%3A%22third_party%22%2C%22listsBrand%22%3Afalse%2C%22filter%22%3A%7B%22brand%22%3A%5B%22%3Cbrd_id%3E%22%5D%2C%22brandRule%22%3A%22not_named%22%2C%22market%22%3A%5B%22%3Cmkt_id%3E%22%5D%7D%7D
Search the index: GET https://parse.gl/api/public/search?q=<words>

Connect

Connect an assistant: https://agent.parse.gl/mcp

An agent can save work before signup: POST https://agent.parse.gl/v1/workspaces

What it costs

No account: Public figures and up to 5 examples per list, about 120 brands a day per address.
Free account: $5 in credit, your brand's first check (20 buyer questions) and a free weekly check.
After your first top-up: Exports and deeper search results. The first top-up starts at $5.
Live checks: $0.01 per question per engine ($0.05 on Gemini and Claude).
Full public answers and pages: 25 free per day, then $0.005 each.
Public answers older than 90 days: $0.001 each the first time your workspace counts or lists one, then yours.

Prices: https://parse.gl/pricing.md

Map of the site: https://parse.gl/llms.txt

Compare which sources cite your brand across ChatGPT, Google AI Mode, and Perplexity

Start free in your browser.
Parse

Index

  • Brands
  • Markets
  • Topics
  • How we measure

For agents

  • Docs
  • Connect
  • Workflows
  • llms.txt

Parse

  • Use cases
  • Pricing
  • Research
  • Blog
  • Compare

Company

  • About
  • Contact
  • Corrections

© 2026 Parse, Inc. All rights reserved.

Privacy PolicyTerms of Service