A recent Tom's Guide article compared ChatGPT, Google Gemini, and Perplexity AI for search capabilities. The problem? These tools serve fundamentally different purposes. Comparing them directly is like judging a scalpel, a Swiss Army knife, and a letter opener in brain surgery. They're not the same tool — and pretending they are produces useless conclusions.
Meaningful evaluation requires appropriate context: comparing tools designed for identical purposes rather than mixing fundamentally different technologies.
The Tools Are Not the Same
ChatGPT functions as a general-purpose assistant — the Swiss Army knife. It can write, reason, code, summarize, and converse. It's not specifically designed for search, and evaluating it as a search engine misses the point of what it is.
Perplexity, by contrast, is a dedicated search tool — the scalpel. It was built from the ground up for retrieval-augmented generation (RAG), meaning it fetches current information from the web and synthesizes it into an answer. That's its entire purpose. Of course it performs better at search than a general-purpose tool.
Gemini sits somewhere between — a generalist with some search integration, but not purpose-built for either direction.
The Absurdity of the Comparison
To illustrate how misguided this kind of comparison is: should we evaluate Excel for spreadsheets and a Ferrari for speed in the same review? Both involve numbers and performance, technically. The categories are incompatible.
Forcing ChatGPT and Gemini into a search benchmark they weren't designed for doesn't reveal anything useful about their capabilities. It just produces misleading rankings that confuse consumers.
The Real Winner — When the Right Question Is Asked
If the question is specifically "which of these tools is best for search," the answer is straightforwardly Perplexity. It's the only tool genuinely designed for that task, with RAG capabilities that give it access to current web information rather than relying on a training cutoff. ChatGPT and Gemini aren't optimized for this, and their performance reflects that — not a fundamental weakness, but a category mismatch.
The takeaway isn't that Perplexity "beats" the others. It's that you should use the right tool for the right job. And tech journalism should stop publishing comparisons that pit fundamentally different things against each other as if the results mean something.