ChatGPT Statistics (2026): 68 Sourced Figures

This page collects figures on how many people use ChatGPT, what they use it for and how the business behind it has grown. Every number links to the document that first published it. Usage and business figures come from OpenAI’s own announcements and its economic research paper. Adoption figures come from Pew Research Center in the United States and Ofcom in the United Kingdom. OpenAI is a private company, so its figures are company-reported and not audited filings. For a comparison of chatbot products for business support and sales, see the StatWharf comparison of AI chatbots.

Key figures at a glance

How many people use ChatGPT

OpenAI reports users as weekly active users. It counts people who used the product at least once in a week. The figures below are company-reported counts, quoted as OpenAI stated them.

The September 2026 figure covers all OpenAI products, not only ChatGPT. An OpenAI post of August 31, 2026 says the free tier keeps ChatGPT available to more than 1 billion weekly active users. In January 2026, OpenAI’s finance chief wrote that weekly and daily active users continue to produce all-time highs. OpenAI did not publish a daily active user count in that post.

OpenAI also compares ChatGPT with rival apps. In March 2026 it stated that ChatGPT has 6x the monthly web visits and mobile sessions of the next largest AI app. OpenAI did not name the data provider behind that comparison in the post.

Message volume

The OpenAI working paper “How People Use ChatGPT” was written by OpenAI’s Economic Research team with Harvard economist David Deming. It was released on September 15, 2025 as a National Bureau of Economic Research working paper. It covers consumer plans only (Free, Plus and Pro). Business, Enterprise and Education plans are excluded.

Daily consumer messages grew close to sixfold between the two June readings in Table 1. OpenAI’s September 2026 post adds a view of individual users. Six months after signup, people on individual plans had tried roughly twice as many distinct tasks as in their first month.

What people use ChatGPT for

The working paper classifies a sample of consumer messages with automated classifiers. No researcher read the messages. The paper groups messages in two ways: by whether they relate to paid work, and by topic.

The paper also sorts messages by the kind of output the user wants. “Asking” means seeking information to inform a decision. “Doing” means asking the model to produce an output such as a draft or a plan. “Expressing” covers messages that share views or feelings without a request.

• Asking made up about 49% of messages.
• Doing made up 40% of messages.
• Expressing made up 11% of messages.

The work share fell because non-work use grew faster, not because work use shrank. Work messages rose in absolute terms over the same year. Coding is a small share of consumer ChatGPT use. The paper contrasts its 4.2% with research on other chatbots, where coding takes a far larger share.

Who uses ChatGPT

OpenAI’s study inferred gender from first names and reported the share of users with typically feminine names. It also compared growth across country income groups.

The paper also reports that nearly half of all messages sent by adults came from users under the age of 26.

ChatGPT adoption in the United States

Pew Research Center surveyed 5,119 US adults from February 17 to 23, 2026. The results were published on June 17, 2026 in “Americans and AI 2026”. Pew asked about ChatGPT each year from 2023. From 2023 to 2025, the question was “Have you ever used ChatGPT?” and was asked only of people who had heard of it. In 2026, the question also asked about other chatbots. Pew marks this as a change in wording, so the 2025 to 2026 change is not a strict like-for-like comparison.

Use differs sharply by age. Adults under 50 use ChatGPT about twice as often as older adults, at 57% against 28%.

Pew also asked about specific chatbots. ChatGPT leads by a wide margin.

Pew’s wider chatbot questions put ChatGPT in context.

• 49% of US adults say they ever use AI chatbots such as ChatGPT, Gemini or Copilot. The share was 33% in 2024.
• 42% of US adults use chatbots to search for information. This is the most common use Pew measured.
• 38% of employed adults use chatbots for tasks at work.

ChatGPT in the United Kingdom

Ofcom’s Online Nation 2025 report was published on December 10, 2025. Its ChatGPT traffic figures use Similarweb data on web visits from UK fixed and mobile devices. Ofcom describes ChatGPT as already the second-largest search service in the UK.

Visits measure traffic, not people. One person can make many visits in a month. The UK figures therefore cannot be compared with Pew’s share of adults or with OpenAI’s weekly user counts.

Business customers and paid seats

OpenAI sells ChatGPT to organizations as well as consumers. It counts any organization that pays for ChatGPT workplace plans or for direct use of its models through the developer platform as a business customer.

In April 2026, OpenAI expected enterprise revenue to equal consumer revenue by the end of 2026. Codex, OpenAI’s coding agent, had 1.6 million weekly users in February 2026, so the April figure nearly doubled it in about six weeks.

Revenue, advertising and funding

OpenAI does not file public financial statements. The figures below come from its own posts. They are stated as annualized run rates or rounded totals rather than audited results.

OpenAI ties revenue to compute. Compute grew from 0.2 GW in 2023 to about 1.9 GW in 2025. Revenue grew about 10x over the same two years. Advertising is the newest revenue line. ChatGPT Ads went from a pilot to a $1 billion run rate in less than 200 days. OpenAI states that ads are labeled, sit apart from answers and do not influence what ChatGPT says.

Method

• Data window. Figures date from 2023 to September 2026. The 2023 and 2024 values serve as a baseline for the trend tables. The page prefers the most recent release of each figure.
• Primary sources only. Each figure links to the body that first published it: OpenAI’s own newsroom posts and its working paper, Pew Research Center, and Ofcom. Statistics aggregators, roundups and articles that re-quote numbers were not used. Third-party user estimates that OpenAI has not confirmed were left out.
• Company-reported figures. OpenAI is a private company. Its user, customer and revenue figures are company-reported and have not been audited or filed with a regulator. Figures stated as “more than” are quoted that way.
• Checks. StatWharf Editorial opened each source in October 2026 and confirmed that the number and its date appear there. Figures without a primary source were dropped. These include reported weekly user figures above 1 billion that appear only in press coverage and statistics sites.
• No modeling. The page reports published figures only. The sixfold growth in daily messages and the near doubling of Codex users are direct comparisons of two published values. No other value is calculated or estimated.

What the figures show

ChatGPT’s reach is now measured in the hundreds of millions each week. OpenAI’s count rose from 700 million weekly users in July 2025 to more than 900 million in February 2026. By September 2026 its products reached more than 1 billion weekly users. Independent survey data points the same way. Pew found that 44% of US adults use ChatGPT, close to twice the share of the next chatbot.

Most use is personal, not work. Nearly three in four consumer messages were unrelated to work in June 2025. Yet the business side is growing fast. More than 40% of OpenAI revenue came from enterprise customers by April 2026. Teams that want a chatbot for customer support or sales can compare products, pricing and fit on the StatWharf comparison of AI chatbots. Other StatWharf data pages are listed in the research directory. Send corrections to any figure to the editorial address below.

Factual corrections are free: editorial@statwharf.com.