OpenAI Says ChatGPT Search Volume Is Roughly One-Fifth of Google’s
ChatGPT consumer product lead Adam Fry says ChatGPT handles roughly one-fifth as many search-like queries as Google. The comparison offers a view into how often people use ChatGPT to seek information, but it is not a direct measure of website visits—and key details about the comparison remain undisclosed.
Fry discussed the figure with Big Technology’s Alex Kantrowitz at an OpenAI-hosted dinner in New York. The transcript, published the following day, describes how OpenAI defines a search: each time a user asks a question that sends ChatGPT to the web counts as one search. If a conversation includes multiple questions that trigger web searches, each is counted separately.
What OpenAI’s measure includes
Fry said early ChatGPT use focused largely on writing and coding tasks that people would not typically have taken to a search engine. Over the past year, he said, users have increasingly asked about people, places and things that require current information, as well as topics such as finance and health. He characterized these as questions users might previously have brought to a traditional search engine.
The figure therefore describes web searches initiated within ChatGPT, not every prompt submitted to the chatbot. It also does not describe how many visitors ChatGPT sends to websites. Those distinctions matter for marketers assessing the service as a discovery channel: query activity and referral traffic measure different parts of the user journey.
Fry did not specify which Google search-volume figure OpenAI used for the comparison, the period covered or the regions included. Without those details, the approximate one-fifth ratio is difficult to reconcile precisely with independent estimates. Big Technology also reported that ChatGPT’s weekly active users had increased 20%, from 1 billion to 1.2 billion, over the preceding two months.
Independent estimates use different definitions
In February, Ahrefs estimated ChatGPT’s search volume at about 12% of Google’s, using a method that differs from OpenAI’s. The analysis started with an estimate of 2.5 billion daily ChatGPT prompts in July 2025, drawn from a Harvard University and OpenAI study. Ahrefs estimated that approximately 65% of those prompts were queries users might previously have taken to Google, then compared the resulting figure with Google’s announcement in March 2025 that it handled more than 5 trillion searches annually.
That 65% share was higher than the proportion of prompts the researchers themselves classified as search-related, according to the account of the analysis. More fundamentally, the methods count different things: Ahrefs classified prompts by the type of request, while Fry’s definition counts occasions when ChatGPT searches the web to answer a question. The resulting estimates should not be treated as interchangeable.
Referral data provides another perspective. Ahrefs’ Web Analytics tracker reported that, in October, Google accounted for 24.74% of tracked traffic across 118,671 sites, compared with 0.36% for ChatGPT. Those figures appeared in the tracker’s Nov. 3 update. They measure visits arriving at websites, rather than the number of search-like queries handled by either service.
Implications for traffic reporting
Google’s parent company, Alphabet, reported search-query growth as a rate rather than a total count in its third-quarter release. CEO Sundar Pichai said AI investments were driving overall search-query growth and described Google’s year-over-year query growth rate as its highest in five years. That statement, like OpenAI’s comparison, does not establish whether ChatGPT use is replacing Google searches.
For marketers planning and evaluating search-related channels, the figures point to a need for clear measurement definitions. Query volume indicates how frequently users seek answers; referral traffic shows how many visits reach a site from a platform. Neither measure alone reveals whether a change in one channel reflects substitution, additional searching or other user behavior.
Ahrefs’ month-by-month referral tracking can provide an ongoing external view of traffic arriving from ChatGPT. OpenAI’s one-fifth comparison may add useful context about search activity within the product, but more detail on its benchmark, time frame and geographic scope would be needed for a closer comparison with other estimates.
