Expect Higher Visibility on Grok. Lower on Perplexity. A Study of Brand Mentions from 17M+ Answers

Key takeaway: You are more likely to find your brand in a Grok answer than any other model. On Perplexity, expect the opposite — it mentions brands less often, and when it does, it mentions fewer of them. That's not a signal about your brand. It's how these models behave.

When we get an answer from an AI model for a question like "What's the cheapest product analytics tool?", we analyze that answer to see what brands the AI model talked about. Gauge extracts all brands mentioned by the AI model in that answer.

In this experiment, we measured the average number of brands that each model talks about across over 17 million answers in the last month. The result: Grok mentioned 9.6 distinct brands per answer on average. Perplexity mentioned 4.1.

The table below shows how many brands each model talked about on average:

Model	Avg distinct brands per answer	% of answers with ≥1 brand
Grok	9.56	95.0%
Google AI Mode	7.69	97.9%
Claude (Anthropic)	7.36	93.6%
Gemini	6.87	94.6%
ChatGPT (OpenAI)	6.53	94.4%
Google AI Overviews	5.77	94.6%
Microsoft Copilot	5.40	94.4%
Perplexity	4.09	82.1%

Perplexity is also the only model that frequently answers without naming any brand at all. In our data, roughly 1 in 5 Perplexity answers contains zero brand mentions. Every other model lands in the 93–98% range.

Where this data comes from

Gauge helps brands improve their visibility and perception by first measuring what models are saying and then optimizing your content strategy. The 17 million+ answers here come from the measuring stage where we run prompts through each AI model on a daily basis. For every answer that comes back, we extract the brands mentioned, sentiment, and citations.

For each answer, we counted the number of distinct brands it mentioned — a brand named three times in one answer counts once — and averaged that across all answers per model. Every prompt runs against every model, so the comparison reflects how the models behave, not differences in what we're asking them. See our docs for how these per-answer extractions roll up into the core AEO metrics like visibility and citation rate.

The avg distinct brands per answer number matters because visibility is a rate. When a model mentions more brands per answer, your chances of being included go up — but so does everyone else's. A high average means the model is simply talking about more brands. Thus, raising visibility percentage across the board. A low average means the model is selective, so getting mentioned at all is a stronger signal.

Perplexity just talks about fewer brands

One explanation for Perplexity's low numbers: it skips brand mentions entirely more than any other model. Remove those zero-brand answers and maybe the gap closes. We ran that test.

Even filtering out answers where no brands were mentioned at all, Perplexity still only averages 4.98 brands per answer compared to Grok's 10.06.

Model	Avg brands per answer (all)	Avg brands per answer (≥1 brand only)
Grok	9.56	10.06
Google AI Mode	7.69	7.86
Claude (Anthropic)	7.36	7.87
Gemini	6.87	7.26
ChatGPT (OpenAI)	6.53	6.92
Google AI Overviews	5.77	6.10
Microsoft Copilot	5.40	5.72
Perplexity	4.09	4.98

What this analysis means for your visibility

Hitting 40% visibility on Grok and 40% on Perplexity are very different achievements. On Grok you're one of nine or ten brands the model squeezed into its answer. On Perplexity you're one of four, and the model only got there after deciding to mention any brand at all, which it declines to do almost 18% of the time.