ChatGPT Is Back Up to 55.5 Percent of Chatbot Web Traffic, but It Held 73 Percent a Year Ago
Similarweb's August figures show ChatGPT regaining share from 52.7 percent three months ago while Gemini slips back to 25.6 percent. Claude grew from 1.9 to 9.3 percent in a year. Apps are not counted.
New Similarweb numbers put ChatGPT at 55.5 percent of AI chatbot website traffic, up from 52.7 percent three months ago. Google’s Gemini, which had been closing the gap, slipped from 27.8 to 25.6 percent. So the short version of the last quarter is that ChatGPT stopped the slide.
The longer version is more interesting. A year ago ChatGPT sat at 73.3 percent. Over twelve months it lost roughly eighteen points of share while Gemini doubled and Anthropic’s Claude grew from 1.9 percent to 9.3 percent, close to a fivefold increase. Behind them the field is thin: DeepSeek at 3.4 percent, Grok at 2.4, Microsoft’s Copilot at 1.6, and Perplexity at 0.9.
One important caveat before anyone draws conclusions. These figures count website visits only. They do not include mobile apps, and they do not include desktop clients. Google routes a great deal of Gemini usage through Android, including push notifications that open the app directly rather than a web page. OpenAI has been pushing its own mobile app and the new ChatGPT Work desktop client. None of that shows up here.
What is behind this
Market share in this business is measured badly, and it is worth understanding why before treating any number as fact. Similarweb estimates web traffic from panel data and public signals. That is a reasonable proxy for a product people mostly opened in a browser tab, which is what chatbots were in 2024. It is a poor proxy for 2026, when the same product ships as a phone app, a desktop app, a sidebar in your email, and an API that other companies build on. A model can be enormously used and barely visible in these charts.
What the trend still tells you, despite the measurement problems, is that the one-winner phase has ended. Claude going from a rounding error to nearly a tenth of web traffic in a year is a real shift, and it happened mostly through developers and work use rather than consumer marketing.
What this means for you: If you are choosing a tool, ignore the leaderboard. The gap between these assistants on everyday tasks is now small enough that habit, price, and which apps you already use matter more than the ranking. If you are running a business, the useful takeaway is that betting everything on one provider looks less obviously safe than it did a year ago, and switching costs are worth thinking about before you are locked in.
Sources
Massachusetts Wants Independent Audits of Frontier Models Every Four Months, and the Labs Disagree About It
Anthropic backs a state bill requiring outside evaluators to check large models for catastrophic risk three times a year. OpenAI and Google want something closer to Illinois's annual audit, warning about a patchwork of state rules.
