ChatGPT still leads most independent measures of AI-assistant reach, with more than 1.1 billion monthly users in May 2026, but Gemini has reached a comparable headline scale and Perplexity's user total looks very different depending on what you count. If you are planning an AEO push and trying to figure out where to put your effort, the honest answer is that no single AI chatbot statistics chart tells you the whole story, so this piece walks through what the real numbers say, engine by engine, and where they disagree.
Who actually has the most users
The cleanest AI search engine comparison available right now comes from Sensor Tower's State of AI 2026 report, covering May 2026 and reported by TechCrunch. It measured five major AI assistants against each other using the same method, which is rare, since most companies report their own numbers their own way.
| Engine | Monthly audience | AI-assistant market share |
|---|---|---|
| ChatGPT | More than 1.1 billion | 46.4% |
| Gemini | 662 million | 27.7% |
| Claude | 245 million | 10.3% |
| Perplexity | Not stated separately | Less than 5% |
| DeepSeek | Not stated separately | Less than 5% |
On that measure, ChatGPT is comfortably ahead. It has almost double Gemini's audience and roughly 46% of ChatGPT market share among the assistants Sensor Tower tracked. That ChatGPT market share figure is the closest thing to a fair, single-method comparison available right now, which is why it anchors the table above.
But that is not the only number in circulation, and it is worth knowing why. Google announced on August 11, 2026 that the Gemini app had passed 1 billion monthly active users. OpenAI, for its part, reported 900 million weekly active users back in February 2026. Those two figures get compared to each other constantly, and they should not be, because one counts people active in a week and the other counts people active in a month. A weekly number and a monthly number are measuring different windows, so putting them side by side and calling it Gemini vs ChatGPT usage is misleading even though both companies published the figures honestly.
The safest way to say it: ChatGPT is still the largest platform on the one independent, apples-to-apples comparison available, and Gemini has grown enough that it can no longer be waved off as a distant second. Neither source proves that Gemini has overtaken ChatGPT, and you should be skeptical of any headline that claims otherwise using two incompatible metrics. Whichever way you read the Gemini vs ChatGPT usage numbers, the gap between the two is now a real strategic question, not a rounding error.
How much smaller Perplexity really is
Perplexity users get talked about a lot for a platform this size, and the reason is that its own reported number is large. The Financial Times reported in April 2026 that Perplexity had passed 100 million monthly active users across its search and agent tools combined. If you only look at that one figure, Perplexity users seem to rival Copilot's, but the independent panels tell a different story once you compare like with like.
That sounds close to Copilot's range. It is not, once you look at the independent panels. Sensor Tower placed Perplexity below 5% of the global AI-assistant market in the same May 2026 comparison where ChatGPT held 46.4%. Similarweb's separate traffic panel called Perplexity the smallest of the four platforms it tracked (ChatGPT, Gemini, Claude, and Perplexity) and said its audience had been declining since an October 2025 peak.
These are not necessarily contradictory numbers. The Financial Times figure counts search and agent tools together, a broader product boundary than a narrow chatbot audience count. Sensor Tower and Similarweb are measuring something closer to core usage. So when you see Perplexity's 100 million next to ChatGPT's 1.1 billion, remember that the two numbers were not built to be compared against each other. Perplexity is real and worth tracking for AEO, but it is a smaller platform than its headline figure suggests, and every independent panel that measured it the same way as the others put it well under 5% share.
Which engines are actually gaining ground
Scale and growth are two different questions, and mixing them up is where most AI search engine comparison pieces go wrong. Similarweb's year-over-year panel, covering June 2025 through May 2026, ranked growth rate this way: Claude first, Perplexity second, Gemini third, ChatGPT fourth.
Claude's share of worldwide generative-AI web traffic roughly quadrupled over that year, the sharpest move of any platform Similarweb tracked, and Claude also posted some of the fastest app growth on US iOS and Android in the same window. That is a real signal. It is also a signal from a small base: Claude sat at 245 million monthly users and about 10% market share in Sensor Tower's May 2026 numbers, well behind ChatGPT and Gemini.
Meanwhile ChatGPT's share of worldwide generative-AI web visits fell from about 76% in June 2025 to roughly 53% in May 2026, according to Similarweb, and Gemini absorbed most of what ChatGPT lost. That decline in share does not mean ChatGPT shrank. Similarweb was explicit that ChatGPT added the most users of any platform in absolute terms during that period, even while growing slowest in percentage terms, because its base was already so much larger than everyone else's. A platform with a billion users needs a much smaller percentage gain to add more real people than a platform with a tenth as many.
So the practical reading for anyone deciding where to focus: ChatGPT is not losing its audience, it is losing exclusivity. The market it used to have almost entirely to itself is now split several ways, and Claude in particular is growing fast enough that it deserves a place in your tracking even though its total reach is still much smaller than ChatGPT's or Gemini's.
Where Copilot and DeepSeek fit
Microsoft Copilot is a genuinely large platform, but its public numbers are unusually hard to compare to the others because Microsoft reports commercial and consumer usage together. The company's FY2025 annual report put Copilot at more than 100 million monthly active users across both categories, and a Microsoft spokesperson told TechCrunch in January 2026 that the combined total had reached 150 million. That is meaningful growth, but the figure blends workplace 365 Copilot use with the consumer chat product, so you cannot isolate how many people are using Copilot the way they would use ChatGPT or Gemini for open search and chat.
DeepSeek's story is different again. The strongest sourced figure available is 22.2 million daily active users in China, reported by Reuters back in January 2025, during DeepSeek's initial spike. That number is old, daily rather than monthly, and specific to one country, so it cannot be turned into a current global ranking next to the other engines here. What can be said with more confidence is that Sensor Tower's May 2026 data kept DeepSeek under 5% of the global AI-assistant market, while also noting that ChatGPT, Gemini, and DeepSeek together accounted for nearly 90% of all time spent across AI-assistant apps in the first quarter of 2026. That combined figure says something about how concentrated attention is at the top of the market, but it does not tell you DeepSeek's individual share, so resist the urge to divide it up and assign DeepSeek a number the source never gave.
Why the charts disagree with each other
If you pull up four different reports on AI chatbot statistics, you will likely get four different rankings, and the reason is not that anyone is lying. Each source measures a different thing.
Statcounter's July 2026 panel, for instance, put ChatGPT at 77.89% of AI-chatbot market share worldwide, a much higher number than Similarweb's roughly 53% web-visit share for the same general period. Statcounter did not explain its denominator in the material available, and that gap between 77.89% and 53% is the clearest illustration of why a number without its measurement method attached is close to useless. A few of the recurring mismatches worth knowing before you quote any figure:
- Weekly versus monthly users. OpenAI's 900 million weekly figure and Google's 1 billion monthly figure are measuring different time windows, so they cannot be ranked against each other directly.
- Users versus visits. Similarweb counted 9.5 billion monthly visits and 655 million unique visitors across generative-AI platforms worldwide. A visit is not a user; one person can generate many visits.
- Core product versus product family. Perplexity's 100 million spans search and agent tools together, and Microsoft's Copilot figure spans commercial and consumer products together, both broader boundaries than a single chatbot app.
- Company-reported versus independently measured. Google and Microsoft report their own totals using their own definitions. Sensor Tower and Similarweb use independent measurement panels that apply one method across every platform, which is why they tend to be the more reliable basis for a real comparison.
None of this means the numbers are unusable. It means every figure needs its source and its date attached, the same way this article has tried to carry them, so a reader can judge what is actually being measured before drawing a conclusion from it.
What this means if you are planning AEO coverage
If you run content or AEO for a portfolio of client brands, the practical takeaway is that a ChatGPT-only visibility plan is now a real gap, not just a theoretical one. ChatGPT still deserves the largest share of attention: it has the biggest audience on every credible panel reviewed here, and Similarweb confirmed it is still adding the most users in absolute terms even as its share slips. But Gemini has grown into a platform you track by default rather than as an afterthought, and Claude's growth rate is fast enough that a brand invisible there could be losing ground somewhere you are not watching. Perplexity and Copilot are smaller by most independent measures, but both have distribution large enough to matter for specific clients, and DeepSeek is worth a periodic check given how concentrated its regional attention has been.
None of this is a case for picking a favorite engine. It is a case for treating "AI search" as several separate platforms with different audiences and different growth curves, and checking each one on its own terms instead of assuming that a ChatGPT citation covers you everywhere your buyers actually are.
