Perplexity
Good for cited web answers on Pro with GPT-6.1 Sol; skip it if you want free help, where Claude scores higher.
Should you use it?
Pick it if
People who want web answers with sources and a choice of top models from several makers, at $17 a month billed yearly.
Best for: Quick web research where you want sources to check each answer (AI apps Top 20)
Skip it if
You want strong free help with hard questions: Claude's free plan scores 80 for research, against 70 here on Pro.
You give up: Trains on chats by default; opting out does not remove data already collected
The short version
- What this is
- An AI answer engine that searches the web and cites its sources. Pro lets you pick models from OpenAI, Anthropic, Google and others for each search.
- Why it matters
- Shows which Perplexity plan unlocks the strong models, what it costs, and how it compares with free options for research.
- Pro ($17/month billed yearly) lets you pick GPT-6.1 Sol, 2nd of 42 on Arena's Search board, which earns a score of 70 for research s1 s2 s3.
- Free names no models and does not let you choose one, so we could not score it s2 s3.
- Chats can train Perplexity's models unless you turn off AI Data Retention s5.
Good at
Weak at
- Free names no models and does not let you choose an advanced one; its plan card says basic models while the FAQ says top models s2 s3.
- Pro and Max are priced only as monthly rates billed yearly, and search limits are not published in numbers s2.
- Chats can be used for training by default, and turning that off does not remove data already collected s5.
Our take
Perplexity's score depends on the plan. Its help page lists, as of 17 October, which search models each plan can pick s3. Pro, at $17 a month billed yearly, includes GPT-6.1 Sol and Claude Sonnet 5.5; Max, at $167 a month billed yearly, adds GPT-6 Astra, Gemini 4 Argon, Claude Opus 5.5 and Claude Fable 5.1 s2 s3. Free uses an automatic "Best" option with a limited number of Pro Searches, and you cannot pick an advanced model s3, so there is no named model to score it on.
Under our rule for every app, Pro is scored on its best model for search answers: GPT-6.1 Sol is 2nd of 42 on Arena's Search board s1. That earns a score of 70 for research. Max's extra models don't rank higher for search (Claude Fable 5.1 is 4th, s1), and at $167 a month Max scores lower on value. The ranks are for high-effort settings, and Perplexity says Thinking is a setting on a model s3, so treat the score as Pro's ceiling.
We checked our reviews for the best non-Anthropic option for research. Among the apps we score, Perplexity Pro's 70 is the highest non-Anthropic score for research; ChatGPT's free plan scores 60. Claude's free plan scores 80 at $0.
Score
A score from 0 to 100 per job, against everything available today. Overall 70 across 1 job, the average of the job scores.
New to this? What the scores and labels mean
- 0 to 100
- Our score for one job, like writing or coding, compared with everything else out today. 100 is the best available today. For AI models we build it from public data, such as Arena's blind votes and Epoch AI's independent tests; apps and tools are scored on our own checklist. The same tool can score differently for different jobs, and where two likely ranges overlap, we call it a tie.
- Our verdict
- Use it: a top choice for most people. Worth trying: good, and worth a go. Depends: good for some people or one kind of job. Wait: too early to tell, or not out for everyone yet. Skip: something else does the job better.
- First look
- Just released. We say what it is and whether to bother, and give it a score once public results exist.
- No score yet
- None of the public sources we use covers it yet, or the only evidence is from a source whose numbers we can't use, so we describe it in words instead.
- Best buy
- Nothing cheaper scores as well for that job.
- Best free
- The best score for that job that costs nothing, through a free plan or by running on your own computer.
- Pay as you go
- No monthly plan: you pay for each use, usually through a developer account.
- Checked
- The day we last confirmed the score and the price against the sources. Prices change often, so check the date.
- AI model
- The engine underneath, like GPT or Claude. You use it through an app or, if you code, directly.
- Open source
- Free to download and run on your own computer, so nothing leaves your machine.
Our checklist score, going by a public leaderboard and the maker’s pages.
Our checklist
1 source so far
Sources and notes
Public leaderboardMaker’s pagesPro (lists GPT-6.1 Sol, Claude Sonnet 5.5, Gemini 3.8 Flash, Kimi K3, GLM 5.3, Grok 4.7 and Nemotron 3.5 Super VL, with no default named, s3): 7/10, does the job 2 (GPT-6.1 Sol (Max), rank 2 of 42 on Arena's Search board with style control, from Arena's open leaderboard dataset published 13 Oct, s1; we use the Search board as the better match for a search app), underlying quality 2, value 1 ($17/month billed yearly, s2), privacy 1 (trains by default, with an opt-out in settings, s5), limits 1 (not published, s2). Max (adds GPT-6 Astra, Gemini 4 Argon, Claude Opus 5.5 and Claude Fable 5.1, s3): 6/10, does the job 2 (Claude Fable 5.1 (High) Search rank 4 of 42 or better, s1), underlying quality 2, value 0 ($167/month billed yearly, s2), privacy 1, limits 1. Free: not scored; Perplexity names no Free models and says you cannot choose an advanced model manually s2 s3. Score from Pro, the higher total.
Alternatives
Better or cheaper for research, from our scores, then the ones we'd also look at. The same kind of product comes first.
- ClaudeBetterApp80FreeScores 80 in research, against 70 here; free. Free plan scores 80 for research at $0, against 70 here on Pro. Side by side
- ChatGPT (GPT-6)App60$20 a monthNext best non-Anthropic app we review for research: 60 on its free plan, with images and voice. Side by side
- Gemini Notebook (formerly NotebookLM)AppFirst look: Worth tryingFreeFree tool that answers from files you upload, when your sources are documents rather than the open web. Side by side
Value for money
Where it sits in research: higher is a better score, further left is cheaper. Green dots are the Best buys.
Evidence
Everything behind the call and the scores, folded so the page stays short. Spotted a wrong price, score or link? Report a mistake.
How we scored this
The research score comes from our own checklist of the product.
Public leaderboardMaker’s pages
Reviews are drafted with help from AI models, including Anthropic's Claude. Model scores come only from outside leaderboards and independent evaluations, by a written rule.
No one paid for this review or gave us free access, and we earn nothing if you sign up. How we score
Sources 5
- s1Arena open leaderboard dataset: search_style_control, latest split (published 2026-10-13), CC BY 4.0
- s2Perplexity pricing
- s3Perplexity Help Center: What advanced AI models are included in my subscription?
- s5Perplexity Help Center: Data collection at Perplexity
- s6Perplexity blog: Series A funding round and iPhone app launch (28 Mar 2023)
Something here out of date or wrong? Report a mistake and we will check it against the source.
History 2 changes
Every change to a rating, the price or our call, with the reason. Scores are against what exists on the day, so they drift down as better things ship. Older entries come from before we moved to scores.
- New: Research ratedReviewed: research, rated from Pro. Ranks re-taken from Arena's open dataset (published 13 Oct) instead of arena.ai's web page, which showed GPT-6.1 Sol 8th the same day; the rating is the same either way. A person confirms the rank against the dataset file before import, because the dataset was read through its web viewer.
- released is 2023-03-28, the iPhone app launch; Perplexity says its answer engine first came out in December 2022 but gives no day. The catalog has 2026.
Use it well 4 guides on this site
Pages on this site that help with the jobs above.