Which AI engine names the most companies, ChatGPT, Claude, or Gemini

If a company can only afford to monitor, or optimize for, one engine, it matters whether the engines behave similarly or differently in how often they name any company at all. This finding reports the per-engine visibility rate: of all companies in the cohort, what share appeared in at least one answer from each specific engine.

Across the study's 170,457 usable answers, ChatGPT named companies at a rate of 42.7% (13,057 companies visible on ChatGPT), Claude at 39.9% (12,206 companies), and Gemini at 33.6% (10,284 companies), out of 30,564 companies in the cohort.

The gap between the highest and lowest, ChatGPT at 42.7% against Gemini at 33.6%, is a real difference, not noise from a small sample; it is measured across tens of thousands of answers per engine within the 170,457-answer dataset. Claude sits between the two, at 39.9%.

Why might one engine name more companies than another? A few mechanisms are consistent with this pattern. An engine that tends to produce longer answers, or answers that list more alternatives before settling on a recommendation, will mechanically name more companies per prompt, which raises its visibility rate without necessarily reflecting a broader underlying knowledge base. An engine that is more conservative, naming only companies it is more confident about, would show a lower rate for the opposite reason: not less knowledge, but a higher bar for inclusion. Retrieval behavior differs too: an engine that leans more heavily on live web retrieval versus trained-in knowledge may surface a different, possibly larger or smaller, set of current companies depending on how its retrieval step is scoped.

This study cannot distinguish 'ChatGPT knows about more companies' from 'ChatGPT's answer style includes more companies per answer' from the visibility rate alone, because both produce the same observable number. What would separate the two explanations is a look at average answer length or average number of companies named per answer across engines, which this dataset's fact table does not break out separately from the visibility rate itself. A reader who wants to test the 'longer answers' explanation would need to check whether average companies-per-answer tracks the same ordering as the visibility rate.

What this means for average mention position is a useful check: despite ChatGPT naming companies more often, its average mention position, 5.25, is close to Claude's 5.38 and Gemini's 4.85, against an overall average of 5.18. So the difference between engines shows up mainly in whether a company gets named at all, not in how prominently it is placed once it is.

For a company deciding where to focus monitoring effort with limited resources, this finding says the engines are not interchangeable proxies for each other. Coverage on ChatGPT is not a safe stand-in for coverage on Gemini, and a company that only checks the engine with the highest baseline naming rate may be overestimating how visible it is on AI assistants generally.