AI: which would you rather ask?

AI pack ranks five assistants by which one you'd rather ask: ChatGPT, Claude, Gemini, Grok and Perplexity. They are shown two at a time, and you tap the one you'd rather ask until every pair is settled. The finished order is your own taste for this session, not a benchmark, not a leaderboard, and nobody pays for a rank.

Questions

Which assistants are in the AI pack?

Five: ChatGPT, Claude, Gemini, Grok and Perplexity. Each is a general-purpose assistant you can ask in plain language, in a chat box, about whatever you are working on. ChatGPT, Claude, Gemini and Grok are usually opened as conversational assistants; Perplexity is usually opened when the ask looks more like a search that should come back with sources. They are shown two at a time and you tap the one you'd rather ask, until every pair has been settled and the pack has your order. The list is fixed for this pack; no model version, no API tier, no pricing plan and no sixth assistant is mixed in, and no product link sits behind a tap.

Do ChatGPT, Claude, Gemini, Grok and Perplexity all do the same job?

They overlap as places you ask in chat, but they are not identical products. ChatGPT, Claude, Gemini and Grok are commonly used as conversational assistants; Perplexity is often reached when the ask looks more like a search with sources. Some lean toward long writing or coding help, some toward short conversational replies, and the mix shifts as each product changes. Underneath, each one runs its own models, its own retrieval and its own tooling, so the same question can come back in a different shape depending on which you opened. This pack does not teach those differences as a review or a buying guide, and it does not test any of them. It only asks which one you would rather open when you have a question, so the ranking is about the habit of asking, not a feature matrix or a price comparison.

How does the AI pack ranking work?

Rank'em shows two assistants at a time under the same prompt: which would you rather ask. You tap one, then the next unsettled pair appears, and so on until every pair is settled. With five items that means a short run of forced choices rather than stars or a score sheet. The site turns those pairwise taps into one ordered list for this session. There is no score to grind, no timer, no undo ladder and no second pass that revises what you already chose. When the last pair is done, the pack is ranked: what you tapped is the ranking.

Is this a benchmark of AI models?

No. Nothing here is measured, timed, graded or A/B tested. There is no accuracy score, no latency table, no capability checklist and no shared exam the five take. ChatGPT, Claude, Gemini, Grok and Perplexity can each answer the same question differently, and this page does not try to decide which answer is better or which model is stronger on coding, search or chat. Benchmarks elsewhere measure tasks; this pack only records preference. The order comes only from which one you tapped in each pair, so it is a taste order for the day, not an evaluation.

Is this a leaderboard, and can anyone pay for a higher rank?

No on both. This is not a public climb and not a crowd table of the best assistant. The finished order is yours for this session from your taps alone. There is no aggregate scoreboard that other people climb, no sponsored slot, and nobody pays this site for placement or for a higher position in the pack. Footer line stays the same: not a leaderboard.

Is my AI ranking saved, shared, or tied to a name or email?

The ranking you finish is for this session on this device. There is no sign-in, no name field and no email field on this page or elsewhere on the site. You are not building a public profile, and the pack does not require an account to play. Optional share is a poster of your own order if you choose to save or send it; that is still your list, not a paid placement and not a sitewide leaderboard. Ranking stays client-side for the session — the page is not collecting prompts, chat history or assistant account details.

Why rank assistants if the page is not picking a winner?

Because Rank'em is a preference exercise, not a verdict. Putting five named assistants in pairs forces a concrete choice about which one you reach for, instead of a vague claim that they are all fine or that one is objectively best. Education here means naming the tools, explaining the pairwise mechanic, and stating clearly that the finished list is yours alone. When the pairs are done you have a personal order you can disagree with tomorrow. The page describes what is being ranked and how the taps work; it does not tell you which assistant to subscribe to, which model is best, or where to buy access.

Does tapping an assistant open that product or an affiliate link?

No. A tap only records which of the two you would rather ask in this Rank'em session. It does not open ChatGPT, Claude, Gemini, Grok or Perplexity, and it does not jump to a signup page, a pricing page or an app store listing. It does not follow an affiliate or paid placement URL, and Amazon product or Associates links are not part of this pack. The page stays education about how the pack works and what is being ranked, not a storefront and not a referral funnel.