For the first time, OpenAI has released country-by-country statistics revealing how people worldwide use ChatGPT. The data—published by the OpenAI Economic Research Team on the OpenAI Signals platform—indicate that ChatGPT is increasingly being used to accomplish tasks, not just to answer questions, and that adoption is expanding beyond early adopters.
What the dataset covers
The Signals dataset reflects messages sent within ChatGPT Free, Go, Plus, and Pro accounts, which generally represent individual rather than organizational users. OpenAI says these country-level statistics are intended to help policymakers, researchers and others better understand how more than 1 billion people put ChatGPT to work.
From asking to doing
The release emphasizes a clear shift from exploratory use toward task-oriented use. Practical applications cited include editing, coding, and analysis. Globally, users are more than twice as likely to use ChatGPT in a task-oriented way at work than outside work, while non-work use remains more exploratory with information-seeking still the largest single category.
Regional adoption and ranking changes (Q1 to Q2 2026)
Continuing a steady worldwide increase documented in June, countries and regions with lower early adoption rates are closing the gap. In the second quarter of 2026, parts of Latin America, Oceania and Africa showed faster growth than other regions. Peru, Uruguay, and Costa Rica improved most in the global per-capita rankings. Adoption also continues to rise in North America and Europe, but some Southern Hemisphere areas are catching up.
OpenAI notes changes in per-capita message-rank were reported for 144 countries: blue indicates an improved rank, brown a decline, and gray no change, with hatching marking excluded or ineligible countries.
Multimedia use rising after Images 2.0
New capabilities affect how people use the product. Since the release of ChatGPT Images 2.0 in April 2026, the share of messages worldwide classified as multimedia rose to 7.8%. Although multimedia still trails leading use cases—practical guidance, writing, and information-seeking—it has been on a steady increase year-to-date.
Within multimedia use there are notable geographic differences: in some Latin American countries such as Brazil and Colombia, more than one in ten messages is categorized as multimedia. The refreshed release includes eligible Q2 multimedia estimates for 126 countries; some countries are outside the release or lack eligible data.
Older users are contributing a larger share of messages
The data also show demographic shifts by age. Over the past year, the share of messages sent by users over 35 increased in almost every country. Year over year, these users now account for a 5 percentage-point larger share of messages than they did 12 months earlier. This analysis covers only users who self-reported their age on the ChatGPT platform.
Country-level details highlight especially rapid growth in the 35+ cohort in several places: for example, France and Czechia saw the share of messages from users aged 35 or older rise by more than 10 percentage points in the last year. Nearly three-quarters of European countries recorded larger-than-average increases for this cohort, while some Southeast Asian countries, such as Singapore, showed the opposite pattern. Overall, six of eight countries in the Southeast Asian region showed small increases in the 35+ share.
OpenAI reports three-month trailing shares of age-classified messages for 111 countries with complete estimates, comparing June 2026 endpoints to Q2 2025 averages.
Where to find the data and methodology
OpenAI Signals is the OpenAI Economic Research Team’s hub for data, research and analysis. The team also publishes regular updates on how individuals and businesses use ChatGPT. The Signals release includes downloadable CSVs and a methodology description for those who want to examine the underlying data.
