“Your rankings are wrong.” Perfect! That means they work.

“Your rankings are wrong!”

It is the kind of criticism you would expect when publishing global AI rankings. And to be honest, if you look at our overall ranking for the first time, that reaction is understandable.

Where is ChatGPT? Why is Google Gemini not leading? And how can Claude rank higher than ChatGPT on investment, while OpenAI has raised significantly more than Anthropic?

These are valid questions. But they are based on an assumption that does not hold: that our overall ranking is simply a popularity list.

It is not.

At RankmyAI, we work with four distinct rankings: traffic, investment, reviews, and an overall ranking. The overall ranking is not a standalone metric, but a composite index that combines the other three. Each metric is normalised and weighted equally. What we measure is not popularity alone, but overall traction: the extent to which an AI tool attracts users, raises capital, and is valued by its users.

If you look at traffic in isolation, the picture aligns with expectations. ChatGPT and Google Gemini dominate. They are, without question, the most widely used AI tools globally. If that is what you want to measure, the traffic ranking provides the answer.

But the moment you combine traffic with other signals, the story changes.

Take ChatGPT. Its traffic is unmatched. Its investment backing is among the strongest in the market. But when we look at user reviews on platforms like Trustpilot, the scores are relatively modest compared to many other tools. Because reviews are part of the model, this affects the overall outcome. Not because something is wrong with the data, but because the model is designed to capture multiple dimensions of traction, including user sentiment.

The more complex issue lies with investment.

Many people compare OpenAI and Anthropic at the company level. And they are right: OpenAI has raised more capital. So how can Claude end up higher than ChatGPT in the investment ranking?

The answer is methodological. Our rankings operate at the level of individual tools, while investment data is only available at the company level. That creates a structural challenge. There is no perfect way to assign company-level investment to individual tools.

In practice, there are three options.

The first is to assign the full company investment to every tool. That would mean that every product from OpenAI, Google, Meta, or Microsoft would automatically dominate the rankings. Even relatively small or niche tools from these companies would appear at the top. In our view, that would distort reality rather than reflect it.

The second option we use is to divide the total investment equally among the tools within a company. This is not a perfect representation of how capital is actually allocated internally, but it yields a more balanced, comparable outcome across the ecosystem.

A third option would be to distribute investment based on proxies such as traffic or estimated product importance. While this may sound more refined, it ultimately introduces another layer of assumptions that cannot be verified. In that sense, it remains equally artificial.

Our choice is therefore not about being perfectly accurate, but about being consistent, transparent, and fair.

This also explains why Claude can currently rank above ChatGPT on the investment dimension. OpenAI has multiple tools, including ChatGPT and Sora, meaning that its total investment is split across these products. Anthropic, with a more concentrated product portfolio, allocates its investment across fewer tools. The outcome is a higher per-tool investment figure.

Recent news suggests that OpenAI will discontinue the standalone Sora app. If that happens, the allocation of investment will shift. In practical terms, more of OpenAI’s total investment would then be attributed to ChatGPT, which would likely improve its position in future rankings.

This illustrates an important point: our rankings are not static. They respond directly to changes in how companies structure and position their products.

This brings us back to the initial criticism.

The issue is not that the rankings are wrong. It is that they measure something different from what many people expect. If you are looking for the most used AI tools, the traffic ranking provides a clear answer. If you want to understand where capital is flowing, the investment ranking tells that story. If you care about user sentiment, the review ranking highlights that dimension.

The overall ranking combines all three and therefore answers a different question: which AI tools show the strongest overall traction across usage, funding, and user sentiment.

There is no single “correct” ranking. There are different lenses, each with their own logic.

At RankmyAI, transparency is central. We are explicit about our methodology, the limitations of the data, and the assumptions we make when data is not available. We are also fully independent. There are no paid placements, no sponsored rankings, and no manual adjustments. Every position is the result of the same model applied consistently across all tools.

And importantly, every month is different. The AI landscape is evolving rapidly, with new tools emerging and others disappearing. Our rankings capture that movement. Tools rise and fall based on measurable changes in traffic, investment, and user feedback.

When people say the rankings feel wrong, they often mean the outcome does not match their expectations.
But expectations are usually based on a single signal.
Usage. Visibility. Hype.

We measure three.

And when those signals diverge, the result becomes less intuitive.
But also more informative.

So if the rankings feel wrong at first glance, perfect.

That is exactly when they start to become useful.