ChatGPT Recommends More Companies Than Claude or Gemini

The finding

Across 30564 companies studied, each engine mentioned a different share:

ChatGPT has the broadest reach, mentioning nearly nine percentage points more companies than Gemini. This gap is not explained by the engines answering different questions; all three were given the same 2500 prompts.

What this means

The engines apply different thresholds when deciding whether to recommend a company. ChatGPT appears more willing to name companies, while Gemini is more selective.

Neither approach is inherently better. A broader recommendation set might help buyers discover less well-known options, but it might also include lower-quality recommendations. A narrower set might be more reliable, but it might miss good companies that lack dominant market presence.

Implications for visibility measurement

A company's measured visibility depends heavily on which engine you measure. A company visible on ChatGPT but not Gemini would appear strong in a ChatGPT-only audit but weak in a Gemini-focused one.

29329 companies (96 percent of those visible anywhere) appear on at least one engine but not all three. This near-total figure reflects how different the engines' recommendation universes are.

Why the gap might exist

Training corpus differences. ChatGPT may have been trained on sources that mention more companies, giving it a larger working vocabulary of company names.

Confidence thresholds. Gemini may require stronger evidence before naming a company, leading to fewer but potentially more confident recommendations.

Response length preferences. If ChatGPT tends to give longer responses with more recommendations, it would mechanically mention more companies than an engine that gives shorter responses.

Position when mentioned

Interestingly, when engines do mention companies, they mention them at similar positions:

Gemini mentions companies slightly earlier on average, consistent with it being more selective: if it names fewer companies, the ones it names may be stronger matches.

Practical takeaway

If your customers primarily use one AI assistant, that assistant's opinion matters more for your business than the others. But if your market includes users across all three platforms, measuring on only one gives an incomplete picture. The 42.7-to-33.6 percent range means engine choice can determine visibility.