Responses that name your brand run 10 to 13 percent slower on two engines, and faster on a third
Two earlier posts split citation count and response length by whether an answer named the target brand. Response time had not been checked against that same split. We joined it, and the pattern is not the same on every engine: two engines slow down when your brand gets named, one speeds up, and one barely moves.
Mention-side responses are slower on Claude and Gemini, faster on ChatGPT
We pulled the response time (latencyMs) on 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) and re-scored each one with the scoring engine to flag whether it named the target brand. That is the same 8,664-response, mention/no-mention split used in the citation-count and response-length posts, joined here to latency instead.
| Engine | Mention-side: median (n) | No-mention: median (n) | Delta |
|---|---|---|---|
| Claude | 25.0s (585) | 22.8s (2,095) | +9.7% |
| Gemini | 11.6s (508) | 10.2s (1,818) | +13.4% |
| ChatGPT | 10.4s (327) | 11.0s (973) | -5.1% |
| Perplexity | 3.8s (491) | 3.6s (1,867) | +5.6% |
The mention-side and no-mention-side counts match citation-count-vs-target-mention's split exactly on every engine, confirming this is the same 8,664-response dataset re-joined on a new field, not a different sample. On Claude and Gemini, a response naming the target brand runs measurably longer than one that does not. ChatGPT runs the other way: naming the brand is associated with a slightly faster response. Perplexity moves a small amount in the same direction as Claude and Gemini, but the gap is under six percent against a 3 to 4 second baseline, closer to noise than to Claude and Gemini's double-digit gaps.
The Claude and Gemini gap holds in every category. ChatGPT's does not.
We re-split each engine by category to check whether one crowded vertical was driving the blended numbers.
| Engine / category | Mention: median (n) | No-mention: median (n) | Delta |
|---|---|---|---|
| Claude, HR software | 25.8s (78) | 21.7s (302) | +18.8% |
| Claude, customer support | 24.8s (139) | 23.2s (431) | +6.7% |
| Claude, project management | 24.8s (368) | 23.0s (1,362) | +7.5% |
| Gemini, HR software | 12.7s (54) | 11.0s (213) | +15.6% |
| Gemini, customer support | 13.5s (115) | 10.6s (338) | +27.1% |
| Gemini, project management | 10.7s (339) | 10.0s (1,267) | +6.8% |
| ChatGPT, HR software | 11.0s (28) | 10.9s (122) | +0.5% |
| ChatGPT, customer support | 11.5s (158) | 11.2s (422) | +2.3% |
| ChatGPT, project management | 9.0s (141) | 10.8s (429) | -16.8% |
| Perplexity, HR software | 3.7s (67) | 3.2s (274) | +16.7% |
| Perplexity, customer support | 3.6s (83) | 3.6s (251) | -0.3% |
| Perplexity, project management | 3.8s (341) | 3.7s (1,342) | +4.2% |
Claude is slower on mention-side responses in all three categories (7 to 19 percent). Gemini is slower on mention-side responses in all three categories too (7 to 27 percent). That consistency is what separates a real effect from a category artifact: neither engine's blended number depends on one vertical carrying it. ChatGPT and Perplexity do not hold a direction. ChatGPT is close to flat in HR and customer support but swings to -16.8% in project management, its largest category cell (141 vs 429). Perplexity swings from +16.7% in HR to -0.3% in customer support. The blended near-zero read for both of those engines is a few categories cancelling out, not a stable non-effect everywhere.
Latency tracks the citation effect on the two engines that have one
Two earlier posts split citation count and response length by this same mention/no-mention flag. Putting all three measures side by side shows where the latency gap fits:
| Engine | Latency | Citation count | Word count |
|---|---|---|---|
| Claude | +9.7% | +13.4% | -0.6% |
| Gemini | +13.4% | +14.7% | -11.5% |
| ChatGPT | -5.1% | -0.1% | -12.0% |
| Perplexity | +5.6% | 0.0% | -1.1% |
All three columns are the mention-side value against the no-mention-side baseline, positive meaning mention-side is higher. citation-count-vs-target-mention found Claude and Gemini pull meaningfully more citations on mention-side responses while ChatGPT and Perplexity stay flat. Latency lines up almost the same way: Claude and Gemini are the two engines with both a real citation increase and a real time increase, and the size of the two effects is close on Gemini (+14.7% citations, +13.4% latency) and in the same range on Claude (+13.4% citations, +9.7% latency). ChatGPT and Perplexity, the two engines with a flat citation count, also show the smallest latency moves, though ChatGPT's is a genuine sign flip rather than a null result.
Word count breaks the pattern differently. response-length-vs-target-mention found ChatGPT and Gemini write meaningfully shorter mention-side answers while Claude and Perplexity stay flat, a different engine pairing than the citation or latency splits. ChatGPT is the odd one out across all three measures: it writes shorter mention-side answers, pulls the same number of citations, and returns faster. Nothing here proves causation, but on ChatGPT the pattern reads as one direction, less to write when the target brand is already the obvious answer, rather than more retrieval work. Gemini shows the opposite combination: mention-side answers are shorter and cite more, with a real time cost attached, consistent with more source-checking packed into less prose.
Consistent with the known baseline
Blending mention-side and no-mention-side back together per engine gives median response times of 23.3s for Claude, 10.5s for Gemini, 10.8s for ChatGPT, and 3.6s for Perplexity across 8,664 responses. Those match response-latency-by-engine's published figures exactly, and the per-engine mention/no-mention response counts match citation-count-vs-target-mention's published counts exactly, confirming this is the same dataset read along a new field, not a new sample.
What to take from it
This is a correlation, not a causal claim, and it runs in the direction you would expect once you already know the citation-count result: on Claude and Gemini, the same responses that pull more sources when they name your brand also take longer to write, which is consistent with extra retrieval work rather than the brand name itself adding time. It is not a reason to expect a faster audit turnaround to predict better visibility, and it is not a lever a client can pull. What it does confirm is that the citation-count effect on Claude and Gemini is not an isolated artifact of one measure: it shows up as more sources, more time, and (on Gemini) less prose all at once, the profile of an engine doing more retrieval work on the responses where your brand made the page.
Find out how ChatGPT, Claude, Gemini, and Perplexity source their answers about your category, and where your brand does and does not make the page they read from.
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-read every real (non-error) response from results.jsonl fresh this session, 8,664 in total, and re-scored each one with the scoring engine's brand-mention detection to flag whether the response named the panel's target brand. Response time is the wall-clock latencyMs recorded by our runner, reported as a median. Latency includes provider load at the time of each call and was not randomized across time of day, so small gaps should be treated as directional. ChatGPT contributes 1,300 non-error responses out of 5,225 calls because of quota errors covered in earlier posts, so its figures describe the calls that succeeded. No development-rail or fixture data is included; all responses came from live engine calls. Data collected June-August 2026.