Perplexity keeps its pricing unusually easy to read. There is a free plan, a Pro plan at $20 a month, and a Max plan at $200 a month, with annual billing bringing Pro to about $16.67 a month and Max to about $166.67 across its $2,000 yearly price.
The tenfold jump between Pro and Max is the whole story. These are not three sizes of the same product so much as two different products: a research tool for individuals, and a heavy-duty agent platform for people who run it all day.
Free: the honest taster
The free plan covers everyday questions properly. Quick searches are unlimited, answers carry citations, and you get a small daily taste of Pro Search, the deeper multi-step mode. File uploads and Spaces, Perplexity’s project folders, are included in limited form.
What free will not give you is depth on demand. Deep Research, the model picker and the generous limits sit behind the paywall, and heavy users will meet the caps quickly. As a way to learn whether cited answers fit how you work, though, it is a genuinely useful tier rather than a demo.
Pro: the researcher’s plan
Pro, at $20 a month or $200 a year, is where the product opens up. Pro Search becomes effectively unlimited for normal use, Deep Research runs open up, and the model picker unlocks, letting you choose between leading models from several labs for each question.
Pro also adds serious file uploads, image generation, access to the Labs creation tools, and a monthly credit toward the Sonar API. Students and educators can get Education Pro at $10 a month with verification, which makes Pro’s feature set the cheapest serious research seat in the market.
Max: the power tier
Max, at $200 a month or $2,000 a year, is aimed at professionals whose work is investigation at volume. It raises limits across the board, unlocks the highest-end models and early feature access, and includes a large monthly allowance of credits for Perplexity Computer, its agent that works across apps and files.
It also unlocks extras like the Model Council, which runs several frontier models on the same question in parallel so you can compare their answers. For a journalist, analyst or founder doing daily deep research, that and the Computer credits are the real purchase; the extra searches are almost incidental.
Which one should you buy?
Start free, and let the caps make the decision. If you hit the Pro Search limit a few times a week, Pro at $20 is one of the easiest subscriptions in AI to justify. If you cannot name the workload that needs it, you do not need Max.
Enterprise and education: the other ladder.
Above the individual plans sit team tiers that change the conversation from features to controls. Enterprise Pro is listed at $40 per seat per month, adding single sign-on, SOC 2 compliance and organisational management, while Enterprise Max, at $325 per seat, piles on the highest limits, more Computer credits and data-retention controls.
Those prices explain Perplexity’s two-front strategy. Individuals are won on the $20 Pro plan’s value; companies are won on governance, because no IT department approves a research tool it cannot administer, whatever the employees think of it.
Education deserves its own note. At $10 a month for verified students and teachers, Education Pro undercuts every comparable research tier in the market, and it is plainly a land-grab: habits formed during a degree tend to survive into a career, and Perplexity would like those habits to be its own.
One caution applies to the whole ladder: allowances move. Deep Research quotas, Labs limits and credit amounts have all been revised as the products evolve, so the structure here, free, then $20, then $200, is the durable part. Check the plan page for current allowances before a big commitment.
Refunds and switching are painless by industry standards: plans are self-serve, cancel in a couple of clicks, and downgrading to free keeps your history and Spaces intact. Perplexity is plainly confident that the product, not the contract, keeps people paying.
Max earns its price only when Perplexity is infrastructure for your job: constant Deep Research, agent tasks through Computer, or multi-model checking on questions where being wrong is expensive. For everyone else, the ladder is honest: free teaches you the habit, Pro feeds it, and Max is there when the habit becomes a profession.
0 Comments