Wikipedia Statistics (2026): 86 Stats on Traffic, Bots and AI

A data roundup on Wikipedia in 2026: how big it is, how many people still read it, how much of its traffic now comes from bots and AI crawlers, who edits it, and why it matters so much to ChatGPT, Gemini, Perplexity and Google’s AI answers. Every figure comes from the Wikimedia Foundation, its public analytics APIs or Pew Research Center, and every one was fetched on October 29, 2026.

The short answer: Wikipedia is still one of the biggest sites on the web, with nearly 15 billion views a month and more than 65 million articles (Wikimedia Foundation). But human readership is shrinking. After Wikimedia rebuilt its bot detection, human pageviews were down roughly 8% against the same months of 2024 (Wikimedia Foundation), and in the July 2026 Movement Metrics report mobile web pageviews were down 16% year on year (Movement Metrics). Meanwhile bots produced more than 88 billion Wikipedia views in 2025 (Pew Research Center). Fewer people click through, but almost every large language model is trained on Wikipedia, which makes it one of the most important sources behind AI answers.

Key takeaways

• Nearly 15 billion views a month and over 65 million articles in more than 300 languages (Wikimedia Foundation)
• Roughly 8% decline in human pageviews against the same months of 2024, which Wikimedia links to generative AI and social media (Wikimedia Foundation)
• 16% year-on-year drop in mobile web pageviews in the July 2026 report, a measure less affected by bots (Movement Metrics)
• More than 88 billion Wikipedia views came from web crawlers, AI bots and other nonhuman agents in 2025 (Pew Research Center)
• At least 65% of Wikimedia’s most resource-heavy traffic comes from bots, even though bots are about 35% of pageviews (Wikimedia Foundation)
• About 1.5 billion crawler requests are blocked or throttled every day (Wikimedia Foundation)
• Almost all large language models train on Wikipedia datasets (Wikimedia Foundation)
• Visibility in AI outputs is now driving promotional editing on Wikidata, according to Wikimedia’s own analysts (Movement Metrics)

How Wikipedia traffic is measured

Wikipedia’s numbers come in a few flavours, and mixing them up is the most common mistake in articles about Wikipedia statistics. The Wikimedia Foundation counts every request for a page, then sorts it by who it thinks made it. Since 2025 it has also revised its history several times as it caught bots that were pretending to be people, so older and newer figures are not always directly comparable.

Definitions used on this page

• User pageviews: page requests that Wikimedia believes came from people. Wikimedia’s own definition is requests that did not come from self-declared bots (“spider”) or traffic it classified as bots (“automated”).
• Spider pageviews: requests from crawlers that identify themselves as bots, such as search engine crawlers.
• Automated pageviews: requests that Wikimedia’s systems classified as bots even though they did not say so.
• Unique devices: an estimate of how many distinct devices visited a project in a month. Wikimedia stopped publishing it in its monthly report after bots distorted it in 2025.
• Active editor: in Wikimedia’s movement metrics, a registered user with at least 5 content edits in a month. The 25th anniversary figure of “nearly 250,000 editors” uses a looser definition: one edit or more in a month.
• Developed Wikipedias: Wikimedia’s term for the 8 largest language editions (Chinese, English, French, German, Italian, Japanese, Russian and Spanish). Everything else, apart from Wikidata and Commons, is an “emerging wiki”.
• Wikimedia Enterprise: the Foundation’s paid product for companies that reuse Wikipedia content at large scale, including AI and search companies.

Where we did our own arithmetic on Wikimedia’s raw API numbers, such as year-on-year changes, we label it as a tools8020 calculation and keep it in tables and prose rather than in the numbered list.

How big is Wikipedia?

1. Over 65 million articles across more than 300 languages, according to the Wikimedia Foundation’s 25th anniversary release (Wikimedia Foundation).
2. More than 67 million articles across over 300 languages, in the Foundation’s October 2026 update (Wikimedia Foundation).
3. Over 66 million articles across all languages as of December 2025 (Pew Research Center).
4. Around 7 million of those articles are in English, the language with the most articles (Pew Research Center).
5. Roughly 775 terabytes of storage hold the text, images, videos and other files behind those articles (Pew Research Center).
6. 7,256,657 content articles on English Wikipedia, from the live site statistics counter (Wikipedia API).
7. 66,499,476 total pages on English Wikipedia across all namespaces, including talk, user and project pages (Wikipedia API).
8. 5,310,242,919 words across English Wikipedia articles, as counted by the site’s search index (Wikipedia API).
9. Over 5 billion words in English Wikipedia, which would take one person about 38 years to read (Pew Research Center).
10. 144 million images, videos and other files sit on Wikimedia Commons, the media library behind Wikipedia (Wikimedia Foundation).
11. 1,377,834,222 edits have been made to English Wikipedia since it launched (Wikipedia API).
12. 55,206,971 registered accounts on English Wikipedia (Wikipedia API).
13. Top-ten website: Wikipedia is the only site among the ten most visited in the world that is run by a nonprofit (Wikimedia Foundation).

By our arithmetic on the live counter, English Wikipedia articles average roughly 730 words each (5.31 billion words over 7.26 million articles). English is the largest edition, but it is only about one in nine of the 65 million-plus articles worldwide.

Source: English Wikipedia siteinfo statistics API, fetched October 29, 2026.

How many people read Wikipedia?

14. Nearly 15 billion views every month, according to the Wikimedia Foundation (Wikimedia Foundation).
15. Up to 15 billion page views per month, in the Foundation’s October 2026 statement (Wikimedia Foundation).
16. Over 1.5 billion unique devices access Wikipedia every month, per the 25th anniversary release (Wikimedia Foundation).
17. 1.9 trillion total views of Wikipedia articles in the past decade (Pew Research Center).
18. About 508 million views per day on average over that decade (Pew Research Center).
19. 49% of all Wikipedia views since 2015 were of English Wikipedia (Pew Research Center).
20. 12,923,604,893 user pageviews across all Wikimedia projects in September 2026 (Wikimedia API).
21. 14,033,875,815 user pageviews across all projects in September 2025 (Wikimedia API).
22. 14,687,879,564 user pageviews across all projects in September 2024 (Wikimedia API).
23. 6,271,394,362 user pageviews on English Wikipedia in September 2026 (Wikimedia API).
24. 6,916,125,844 user pageviews on English Wikipedia in September 2025 (Wikimedia API).
25. 3,841,437,955 mobile web user pageviews on English Wikipedia in September 2026, against 4,423,983,926 in September 2025 (Wikimedia API).
26. 2,267,771,569 desktop user pageviews on English Wikipedia in September 2026, against 2,320,019,106 in September 2025 (Wikimedia API).
27. 768,028,180 unique devices visited English Wikipedia in September 2026, against 992,971,201 in September 2025 (Wikimedia API).

The raw API numbers tell a consistent story. Every reader measure we pulled is lower in September 2026 than a year earlier, and mobile web, where most human reading happens, fell faster than desktop.

Source: tools8020 calculation from the Wikimedia Analytics REST API (pageviews and unique devices), fetched October 29, 2026.

Over two years the drop is bigger. All-projects user pageviews fell from about 14.69 billion in September 2024 to 12.92 billion in September 2026, a fall of about 12.0% by our arithmetic. Treat the unique devices figure with care: Wikimedia excluded it from its own monthly report after bots inflated it in 2025, so the September 2025 baseline may still contain some bot traffic.

Is AI reducing Wikipedia traffic?

Wikimedia says yes, at least in part. In October 2025 it published the first official estimate after reworking its bot detection, and its monthly reports since then show the decline continuing.

28. Roughly 8% fewer human pageviews than in the same months of 2024, after Wikimedia reclassified its traffic (Wikimedia Foundation).
29. March-August 2025 traffic was reclassified after Wikimedia found that much of an unusual spike came from bots built to evade detection (Wikimedia Foundation).
30. About 4% year-on-year fall in user pageviews in the July 2026 Movement Metrics report, which Wikimedia says substantially understates the true decline because of undetected bots (Movement Metrics).
31. 16% year-on-year fall in mobile web pageviews in the same report, a measure Wikimedia says is less affected by bots (Movement Metrics).
32. May 2025 is when Wikimedia stopped including unique devices in its monthly report, because bot traffic had distorted them so heavily (Movement Metrics).

Wikimedia’s explanation is direct. It wrote that it believes the declines “reflect the impact of generative AI and social media on how people seek information”, that search engines increasingly answer questions with generative AI instead of linking out, and that younger people look for information on social video platforms (Wikimedia Foundation). Pew Research Center, in its 25th anniversary analysis, also pointed to that roughly 8% fall (Pew Research Center).

This matters for anyone who depends on organic traffic. Wikipedia ranks for an enormous share of informational searches, so it is an early signal of what AI answers do to click-through across the web. Our AI search statistics, Google AI Overviews statistics and Google AI Mode statistics cover the same shift from the search engine side.

How much Wikipedia traffic comes from bots?

33. More than 88 billion Wikipedia views in 2025 came from web crawlers, AI bots and other nonhuman agents (Pew Research Center).
34. 7,169,954,906 automated (undeclared bot) pageviews across all Wikimedia projects in September 2026 (Wikimedia API).
35. 6,082,474,546 automated pageviews across all projects in September 2025 (Wikimedia API).
36. 6,110,936,569 spider pageviews, from self-declared crawlers, across all projects in September 2026 (Wikimedia API).
37. About 35% of total pageviews come from bots, by Wikimedia’s April 2025 measure (Wikimedia Foundation).
38. At least 65% of the most resource-consuming traffic to Wikimedia’s core datacenters comes from bots (Wikimedia Foundation).
39. 50% growth in bandwidth used to download multimedia since January 2024, driven largely by scrapers feeding images to AI models (Wikimedia Foundation).
40. About 30% of all automated requests from crawlers that ignore Wikimedia’s policies are now blocked or throttled (Wikimedia Foundation).
41. From 25 to 30%: Wikimedia revised that blocked share upward in an April 15 update to include blocked media requests (Wikimedia Foundation).
42. About 1.5 billion requests are blocked or throttled every day (Wikimedia Foundation).
43. Approximately 300 million additional unique devices, mostly geolocated to Brazil, appeared in May 2025 because of bots that bypassed detection (Wikimedia data issue log).
44. -15.5%: correcting for those bots cut all-projects user pageviews for May 2025 from 17.2B to 14.5B (Wikimedia data issue log).
45. -14.3%: the same correction cut June 2025 user pageviews from 15.5B to 13.3B (Wikimedia data issue log).
46. -18.9%: unique devices for all Wikipedias in June 2025 were revised from 2.08B to 1.69B (Wikimedia data issue log).
47. Millions of automated requests to Wikimedia’s public APIs came from agents Wikimedia believes were operated by OpenAI, according to an October 2026 investigation (Wikimedia Foundation).
48. Hundreds of thousands of Wikidata Query Service queries came from the same agents (Wikimedia Foundation).

Put September 2026’s three traffic types side by side and the human share is now under half of all recorded pageviews.

Source: tools8020 calculation from Wikimedia Analytics REST API pageviews by agent type (all projects), fetched October 29, 2026.

Automated pageviews rose about 17.9% year on year, from about 6.08 billion in September 2025 to 7.17 billion in September 2026, by our arithmetic. These figures only count page requests Wikimedia recorded. Blocked requests, API calls and media downloads, where the AI scraping load is heaviest, are measured separately. For the wider picture of how AI crawlers hit publishers, see our AI crawler statistics.

Who edits Wikipedia?

49. Nearly 250,000 editors work on Wikipedia every month, counting anyone with a username who makes one edit or more (Wikimedia Foundation).
50. 324 edits per minute across Wikipedia, per the 25th anniversary release (Wikimedia Foundation).
51. More than 340 edits every minute, in the Foundation’s 2024-2025 annual report (WMF annual report).
52. 5.7 edits every second on average in 2024 (Pew Research Center).
53. Nearly 40% of those edits came from Wikipedia-approved bots that fix links, update statistics and revert vandalism (Pew Research Center).
54. More than 600,000 active users across Wikipedias, meaning account holders with at least one edit in the last 30 days (Pew Research Center).
55. 45% of those active users contribute to English Wikipedia (Pew Research Center).
56. More than 15 million registered English users have ever made an edit (Pew Research Center).
57. 270,134 active users on English Wikipedia on October 29, 2026 (Wikipedia API).
58. 818 administrators on English Wikipedia (Wikipedia API).
59. 90,585 registered users made content edits to English Wikipedia in September 2026, against 91,026 in September 2025 (Wikimedia API).
60. 5,196,974 edits to English Wikipedia in September 2026, against 5,272,619 in September 2025 (Wikimedia API).
61. 16,461 new content pages created on English Wikipedia in September 2026, against 19,822 in September 2025 (Wikimedia API).
62. 11% year-on-year drop in retained newcomers at the 8 largest Wikipedias in July 2026; English Wikipedia was flat (Movement Metrics).
63. 14% drop in retained newcomers at emerging wikis (Movement Metrics).
64. 10% year-on-year fall in desktop account registrations (Movement Metrics).
65. Roughly double: mobile web registrations reached about twice their March level and overtook desktop for the first time (Movement Metrics).
66. 0.7% year-on-year decline in active administrators at the 8 largest Wikipedias, and 7.3% at emerging wikis (Movement Metrics).

The editor base is holding up better than readership. English Wikipedia’s monthly content editors were almost flat year on year (about -0.5% by our arithmetic), and edits were down about 1.4%. New article creation fell harder, by about 17%.

Source: Wikimedia Analytics REST API (editors, edits and new pages), fetched October 29, 2026.

What do people read on Wikipedia?

67. Nearly 300 million views made Donald Trump’s article the most-viewed English page of the last decade (Pew Research Center).
68. Nearly 15 million visits to the Charlie Kirk article on Sept. 10, 2025, the most for any article in a single day (Pew Research Center).
69. Nearly half a billion views went to the “Deaths in [year]” pages for 2015-2025 (Pew Research Center).
70. 46% of each month’s most-viewed English articles since 2015 were about prominent people (Pew Research Center).
71. 14% of those top monthly articles were events such as the FIFA World Cup, and 31% were movies or TV shows (Pew Research Center).
72. More than 59,000 revisions make the List of WWE personnel the most-edited English article (Pew Research Center).
73. 2,857 edits in 24 hours to the article on the July 7, 2005 London bombings, the single-day record (Pew Research Center).
74. More than 2.8 million views of Jimmy Carter’s page in a day after his death in December 2024 (Wikimedia Foundation).
75. 8,522,591 views of the Lizzie Borden article in September 2026, the most-viewed English article that month once the Main Page and special pages are excluded (Wikimedia API).
76. 204,438,595 views of the English Main Page in September 2026 (Wikimedia API).

People still use Wikipedia mostly to check who someone is and what just happened. That is exactly the kind of question AI assistants now answer directly, which helps explain why reader numbers are slipping.

How do AI companies use Wikipedia?

77. Almost all large language models (LLMs) train on Wikipedia datasets, according to the Wikimedia Foundation (Wikimedia Foundation).
78. Mistral AI, Perplexity, Ecosia, Microsoft, Pleias and ProRata became new Wikimedia Enterprise partners over the past year (Wikimedia Foundation).
79. Amazon, Google and Meta were already Wikimedia Enterprise partners (Wikimedia Foundation).
80. Visibility in AI outputs is the likely motive behind a jump in promotional editing on Wikidata, which has kept Wikidata registrations dramatically elevated (Movement Metrics).
81. Hundreds of thousands of data queries and millions of crawled pages, mainly from Wikidata and Wikimedia Commons, came from AI agents Wikimedia attributes to OpenAI (Wikimedia Foundation).

Wikimedia Enterprise is how AI and search companies license Wikipedia at scale. The 2026 partner list now includes an AI answer engine (Perplexity), a model lab (Mistral AI) and Microsoft, alongside Amazon, Google and Meta (Wikimedia Foundation).

The Wikidata finding is the one marketers should pay most attention to. Wikimedia’s own analysts say people are now editing Wikidata, the structured database behind Wikipedia’s facts, because they believe it will raise their visibility in AI answers (Movement Metrics). That is promotional editing, which breaks Wikimedia’s rules and tends to be reverted. The better lesson is that AI models lean on trusted third-party references, so what Wikipedia and other independent sources say about your category shapes what ChatGPT and Gemini say about you.

Why Wikipedia matters for AI visibility

Wikipedia sits at both ends of the AI search shift. On one side, it is losing human readers as AI answers replace clicks. On the other, it is one of the most important inputs to those AI answers, because nearly every major model is trained on it and AI search engines retrieve it.

For brands, that has three practical consequences:

1. Your category’s Wikipedia coverage shapes AI answers. If the article on your product category describes the market, lists the leaders or defines the terms, models repeat that framing. You cannot buy or write your way into it, but you can make sure accurate, independent sources about you exist for editors to cite.
2. Reference-style content gets cited. AI engines favour sources that read like Wikipedia: clear definitions, numbers with sources and neutral comparisons. Our GEO statistics show that adding citations, quotes and statistics lifts AI visibility the most.
3. Measure what AI engines actually say. Wikipedia’s traffic drop shows clicks are a weaker signal than they used to be. You need to see whether ChatGPT, Perplexity, Gemini and Google’s AI answers mention you and which sources they cite.

That last point is what AI visibility tools are for. Mentionova tracks a set of prompts across ChatGPT, Perplexity, Google AI, Claude and Gemini, shows which sources each engine cites (Wikipedia included) and flags answers that get your pricing or features wrong. For a wider shortlist, see our best AI visibility tools, best AI citation tracking tools and best GEO tools. Reddit plays a similar role as a source, which our Reddit statistics cover, and our ChatGPT statistics show how big the AI answer audience has become.

Stats ledger: what each source measured

Source: tools8020 stats ledger for this page. Every row links to the primary page or API we fetched on October 29, 2026.

What to do with this data

1. Stop treating Wikipedia traffic as a stable benchmark. Human pageviews are falling and the counts get revised as bots are caught. Use Wikimedia’s latest corrected numbers.
2. Expect fewer clicks from informational queries. If a site as dominant as Wikipedia is losing readers to AI answers, your top-of-funnel content will too.
3. Be cited, not just ranked. AI answers draw on reference sources. Write pages with clear definitions, sourced numbers and fair comparisons.
4. Get covered by independent sources. Wikipedia editors cite news, books and research, not company blogs. Earned coverage is what makes a brand citable, both on Wikipedia and in AI answers.
5. Don’t edit Wikipedia or Wikidata to promote yourself. Wikimedia is already watching promotional editing aimed at AI visibility.
6. Watch your crawler bills too. If bots are 65% of Wikimedia’s most expensive traffic, check your own logs for AI crawlers.
7. Track AI answers directly. Use an AI visibility tool such as Mentionova to see how ChatGPT, Perplexity, Gemini and Google’s AI describe your brand and which sources they lean on.

Frequently asked questions

How many people visit Wikipedia each month?

The Wikimedia Foundation says Wikipedia is viewed nearly 15 billion times a month and accessed by over 1.5 billion unique devices. Its own API recorded 12,923,604,893 human (user) pageviews across all Wikimedia projects in September 2026.

How many articles does Wikipedia have?

More than 65 million across more than 300 languages, by the Foundation’s count, rising to more than 67 million in its October 2026 statement. English Wikipedia had 7,256,657 content articles on October 29, 2026.

Is Wikipedia traffic declining?

Yes. After reclassifying bot traffic, Wikimedia found human pageviews were down roughly 8% against the same months of 2024. Its July 2026 report showed mobile web pageviews down 16% year on year.

Why is Wikipedia losing traffic?

Wikimedia believes the decline reflects generative AI and social media. Search engines increasingly answer questions directly with AI instead of linking out, and younger people look for information on social video platforms.

How much of Wikipedia’s traffic is bots?

Bots were about 35% of pageviews in Wikimedia’s April 2025 measure, but at least 65% of its most resource-heavy traffic. Pew counted more than 88 billion bot views in 2025. In September 2026, automated and spider traffic together made up just over half of recorded pageviews, by our arithmetic on Wikimedia’s API.

Do AI models use Wikipedia?

Yes. Wikimedia says almost all large language models train on Wikipedia datasets, and search engines and social platforms prioritise its information. AI companies including Microsoft, Mistral AI and Perplexity license it through Wikimedia Enterprise.

Which AI companies pay Wikipedia?

The Foundation names Ecosia, Microsoft, Mistral AI, Perplexity, Pleias and ProRata as new Wikimedia Enterprise partners, joining Amazon, Google and Meta.

How many people edit Wikipedia?

Nearly 250,000 editors a month, counting anyone with a username who makes one edit or more. Pew counts more than 600,000 active users across all Wikipedias, 45% of them on English Wikipedia.

What is the most viewed Wikipedia page?

Over the last decade, Donald Trump’s article, with nearly 300 million views. The single-day record belongs to Charlie Kirk’s article, with nearly 15 million visits on Sept. 10, 2025.

How many languages is Wikipedia in?

More than 300, according to the Wikimedia Foundation. English is the largest edition, with around 7 million articles by Pew’s count and 7,256,657 on the live counter on October 29, 2026.

Does being on Wikipedia help you show up in ChatGPT?

It can, because models are trained on Wikipedia and AI search engines cite it. But promotional editing breaks Wikipedia’s rules and gets reverted. The reliable route is earning independent coverage that editors can cite, then tracking how AI engines describe you with an AI visibility tool.

How we compiled these statistics

We fetched 21 primary source pages and API endpoints on October 29, 2026 and recorded 86 data points in docs/content/stats-ledgers/wikipedia-statistics.json. Every numbered figure above maps to a ledger row with a verbatim quote from the page or API response we fetched. Sources are the Wikimedia Foundation (press releases, Diff blog posts, the 2024-2025 annual report and the Movement Metrics report), Wikimedia’s public analytics APIs and the English Wikipedia statistics API, plus Pew Research Center’s analysis of Wikimedia data. We did not use third-party traffic estimators. Year-on-year changes and traffic shares are our own arithmetic on Wikimedia’s raw numbers and are labelled as such. Wikimedia revises historical traffic as it improves bot detection, so later fetches of the same API can return slightly different numbers.

Sources

• Wikimedia Foundation: Wikipedia celebrates 25 years (January 15, 2026)
• Wikimedia Foundation: OpenAI “rogue” agent activities found on Wikimedia projects (October 5, 2026)
• Wikimedia Foundation (Diff): New user trends on Wikipedia (October 17, 2025)
• Wikimedia Foundation (Diff): How crawlers impact the operations of the Wikimedia projects (April 1, 2025)
• Wikimedia Foundation (Diff): Quo vadis, crawlers? (March 26, 2026)
• Wikimedia Wikitech: May 2025 spike in bot traffic
• Wikimedia Movement Metrics, July 2026 report
• Wikimedia Foundation 2024-2025 Annual Report (PDF)
• Pew Research Center: Wikipedia at 25, what the data tells us (January 13, 2026)
• English Wikipedia siteinfo statistics API
• Wikimedia API: all-projects user pageviews, monthly
• Wikimedia API: all-projects automated pageviews, monthly
• Wikimedia API: all-projects spider pageviews, monthly
• Wikimedia API: English Wikipedia user pageviews, monthly
• Wikimedia API: English Wikipedia mobile web pageviews
• Wikimedia API: English Wikipedia desktop pageviews
• Wikimedia API: English Wikipedia unique devices
• Wikimedia API: English Wikipedia editors
• Wikimedia API: English Wikipedia edits
• Wikimedia API: English Wikipedia new pages
• Wikimedia API: English Wikipedia top articles, September 2026