Claude and Gemini cite more sources when they name your brand. ChatGPT and Perplexity do not care.
An earlier post on this blog established that citation count per response is an engine trait, not a category trait: Gemini averages more than double Claude's citation count regardless of what you ask about. That left one question open: within a single engine, does a response that ends up naming your brand cite more sources than one that does not, or fewer? We split all 8,664 real responses in our dataset by whether the target brand was mentioned and recomputed the average citation count on each side, per engine.
Two engines show a real gap. Two show none.
We pulled the citation count from every real, non-error response across our 270 live audit panels (project management, customer support, and HR software; ChatGPT with web search, Gemini with grounding, Perplexity Sonar, and Claude with web search), split by whether that response's text named the target brand. Sample sizes are solid on both sides of the split for every engine, from 327 up to 2,095 responses per cell.
| Engine | Brand named: avg citations (n) | Brand absent: avg citations (n) | Difference |
|---|---|---|---|
| Gemini | 15.3 (508) | 13.3 (1,818) | +14.7% |
| Claude | 6.6 (585) | 5.9 (2,095) | +13.4% |
| ChatGPT | 6.5 (327) | 6.5 (973) | -0.1% |
| Perplexity | 8.8 (491) | 8.8 (1,867) | 0.0% |
On Claude and Gemini, a response that names the target brand carries 13 to 15 percent more citations than one that does not. The median tells the same story: Gemini's median jumps from 11 citations on a no-mention response to 15 on a mention response, and Claude's from 6 to 7. On ChatGPT and Perplexity, the two sides of the split are effectively identical. Perplexity's average does not move at all, consistent with the fixed 5-to-10 citation window it runs on every response regardless of content, documented in the citation-count-per-response post. ChatGPT's average moves by a tenth of a point, noise, not a pattern.
What a wider source pull on Claude and Gemini might mean
This is a correlation in the data, not proof that citing more sources causes a brand to be named, or that being named causes more sources to get pulled. Both engines could be doing either: a response that happens to read a wider set of sources has more chances for one of them to mention your brand by name, or a query that naturally surfaces your brand (because you already rank well for it) also naturally surfaces more of the pages that discuss it. Either mechanism points the same direction for a visibility strategy: on Claude and Gemini specifically, showing up in a response correlates with the engine having pulled a broader source list for that answer, not a narrower one.
Consistent with the known citation-count baseline
The overall averages here (9.5 citations on responses that name the target, 8.8 on responses that do not, 1,911 and 6,753 responses respectively, 8,664 total) match the 8,664-response, 8.9-average baseline from the citation-count-per-response post exactly, confirming this is the same dataset split a different way, not a new or inconsistent sample.
Find out which sources actually get pulled when an AI engine answers a question about your category, and whether your brand's pages are among them.
Get an AI Visibility Audit, $490Methodology
Data drawn from 270 live, search-grounded audit panels (project management, customer support, and HR software brands), each run across four AI engines: ChatGPT with web search, Gemini with grounding, Perplexity Sonar, and Claude with web search. We re-scored every real (non-error) response via engine/scorer.ts this session, both branded and organic prompts, 8,664 responses total, and split the citedSources array length by whether that response's analysis flagged the target brand as mentioned. No development-rail or fixture data is included; all responses came from live engine calls. Data collected June-August 2026.