Perplexity Pricing (2026): Free Is Enough, Max Rarely Is

Perplexity costs $0, $20, or $200 a month for individuals. Free is enough for casual source-backed answers, Pro is the sensible upgrade for recurring research, and Max rarely pays on credits alone: Pro plus 10,000 purchased credits costs $120, still $80 below Max.

Perplexity pricing at a glance

Perplexity has one useful free plan, two paid individual tiers, a half-price education offer, and two seat-based Enterprise tiers. The sticker prices are simple. Computer credits, annual commitments, and separately billed API usage are where the buying decision gets expensive.

The prices and limits below were verified against Perplexity's live pricing page, plan comparison, billing pages, and API documentation on 21 October 2026.

The shortest buying rule is this: stay on Free until the daily advanced-search ceiling interrupts real work, then buy Pro. Treat Max as a capacity plan, not as a better-value Pro. Move to Enterprise for governance, central billing, internal knowledge, and a no-training data policy, not merely because several people need accounts.

The annual rates all save the equivalent of two monthly payments. Pro falls from $240 across twelve monthly charges to $200 upfront. Max falls from $2,400 to $2,000. Enterprise Pro falls from $480 to $400 per seat, and Enterprise Max falls from $3,900 to $3,250 per seat. That is a 16.7% discount against paying monthly.

The free plan is enough for casual research

Perplexity Standard Free is the right plan if your job is asking current questions and checking the sources, not running a daily research operation. It costs $0 and includes practically unlimited basic searches, 3 Pro Searches per day, 1 Research query per month, limited basic file uploads, and automatic model routing, according to the plan comparison.

The boundaries matter more than the word "free." Standard does not provide manual access to advanced models or image generation. It also gives you only a small taste of the heavier modes. Three advanced searches a day can cover a founder checking a competitor announcement, an operator confirming a policy, and one deeper follow-up. It will not support a consultant producing several source-heavy client briefs before lunch.

The paid upgrade must now clear a higher bar because ChatGPT Free offers unlimited text chats with GPT-6, subject to abuse guardrails. OpenAI still limits tool messages, uploads, image generation, voice, deep research, memory, Codex, and Work on the free tier, but ordinary text conversation no longer creates a reason to pay by itself.

That changes the Perplexity value test. Paying $20 merely for "more AI answers" is poor value. Perplexity Pro earns the fee when live sources, citations, Research, files, preferred-model selection, and deliverable creation belong to the recurring job. If the job is casual conversation or an occasional cited lookup, both products have capable free paths.

Who never needs to pay? A solo operator who uses Perplexity a few times a week, stays within 3 Pro Searches a day, needs no more than 1 Research report a month, and does not care which advanced model handles the query can stay on Standard indefinitely. The right upgrade signal is repeated interruption: the allowance blocks paid work, a required model is unavailable, or a file-heavy research workflow keeps moving to another product.

Pro is the sensible individual upgrade

Perplexity Pro is the plan most paying individuals should choose. It costs $20 monthly or $200 annually, expands Pro Search and Research, adds preferred-model selection, supports richer file work, and opens Create files and apps. The annual page rounds the effective rate to $17 a month; the exact $200 charge works out to $16.67.

The payback bar is low when research is part of paid work. At a hypothetical value of $50 an hour, the monthly plan needs to save 24 minutes a month to cover $20. At $100 an hour, it needs to save 12 minutes. One source-backed vendor comparison or market scan can clear that threshold. A person who only asks quick questions may never clear it because Standard already handles the core search job.

Pro's strongest value is workflow compression. A research question can move from current web search to citations, follow-up analysis, file context, and a document or simple app without changing subscriptions. For a funded founder, that can mean turning a market question into an evidence file. For a senior operator, it can mean preserving the source trail behind a recommendation. For a technical buyer, it can mean comparing current documentation without treating model memory as a database.

The hidden limitation is capacity ambiguity. Perplexity says Pro supports a high volume of daily Pro Searches, while its comparison table describes weekly limits for average use. It also warns that advanced-model queries may be limited during especially heavy weeks. Best mode, which lets Perplexity route the query automatically, is available without quota limits. Preferred-model access is not the same as a contractual promise of unlimited use. Perplexity's Pro guide describes those limits.

That ambiguity is acceptable for individual knowledge work and risky for a fixed production schedule. A consultant promising a set number of research reports should check the live account meter before selling the capacity. A mid-market team should not infer an operational service level from the $20 consumer price.

The annual plan is worth taking only after the workflow survives a month or two. It saves $40 a year, but it converts a reversible $20 decision into a $200 upfront commitment. Because the monthly plan already breaks even with minutes of useful work, there is little reason to lock in before confirming that Perplexity, rather than another research tool, is the surface you keep returning to.

Education Pro is the best individual deal

Education Pro costs $10 a month for students and educators who verify through SheerID. It includes everything in Pro plus Learn Mode and education-specific guidance. Perplexity does not publicly list an annual Education Pro rate, so an eligible buyer should budget $120 across twelve monthly charges unless the verified checkout offers something different.

The 50% monthly discount makes Education Pro the obvious paid tier for an eligible researcher, teacher, or student who genuinely needs the Pro workflow. Verification is the boundary. It is not a reusable coupon for alumni, family members, or an organization purchasing seats.

Max is a limits purchase, not a credit bargain

Perplexity Max costs $200 a month or $2,000 a year, ten times the Pro sticker price. It adds the highest access to frontier models, higher Research and creation limits, priority access, Brain in Research Preview, and 10,000 Computer credits each month. Max earns its price only when those higher limits and Max-only capabilities matter together.

The credit math alone does not justify it. Perplexity values 100 Computer credits at $1, so 10,000 monthly credits have a face value of $100. A Pro subscriber can pay $20 for Pro and buy the same 10,000 credits for $100. The combined cost is $120, leaving Max with an $80 monthly premium.

The annual comparison reaches the same result. Annual Pro costs $200. Buying 120,000 credits across the year costs $1,200. The combined $1,400 is still $600 below annual Max at $2,000. There is no steady-state credit-only break-even because both plans buy additional credits at the same published exchange rate. With monthly billing, Max remains $80 more each month after matching the credit balance. This compares credits, not total task capacity: eligible paid Pro accounts also get weekly included usage before credits are charged.

The first month is different because promotional bonuses distort the economics. The credit guide lists a one-time 4,000-credit Pro bonus and a one-time 35,000-credit Max bonus, both expiring in about 30 days. With its normal 10,000 monthly credits, Max starts with 45,000 credits when the bonus is granted. Check the account's exact bonus expiry before planning work around it.

There is no dependable first-month task threshold for upgrading. Pro's included weekly usage is separate from credits and has no published numerical allowance, so compare the credit spend left after that usage with the Max bonus you can use before it expires. A large credit counter is not a forecast of useful output.

Any first-month advantage is temporary. The 35,000-credit Max bonus expires in about 30 days and does not repeat when you cancel, resubscribe, change billing cadence, or apply a later offer. A buyer who cannot use the bonus before expiry receives no residual value. Purchased credits, by contrast, are consumed later and expire only after one year of Computer inactivity.

The practical decision is therefore not "How many credits are included?" It is "Which Max-only limits will be used every month, and what are they worth after the promotional month?" A research professional who regularly hits Pro's advanced-model or Research ceilings may have an answer. A buyer attracted by a large bonus counter does not.

Max annual billing is available only through the web app. It saves $400 against twelve monthly payments, but the safer path is one monthly Max cycle with measured usage. Record Computer credits consumed, Research interruptions, Max-only model usage, and the value of Brain or priority access. If those benefits do not repay $80 beyond equivalent Pro plus credits, return to Pro.

Computer turns the subscription into metered spend

Perplexity Computer is an agent for multi-step work, and its credit billing is separate from ordinary Ask searches. Ask searches consume no Computer credits. Computer's credit cost depends on task complexity, with eligible paid Pro accounts using their included weekly usage first. Perplexity's credit guide explains that boundary.

Since October 20, 2026, Pro, Max, and Enterprise subscribers can run Computer fully locally on supported Apple silicon Macs using Qwen 3.6 or PPLX Pro, Perplexity's post-trained Qwen model. Perplexity recommends 36 GB or more for fully local work; its Hybrid Compute announcement specifies at least 24 GB and macOS 15+ for Hybrid. Fully local keeps planning, reasoning, and file processing on the Mac; Hybrid starts in the cloud, and local tasks can escalate there with permission. For a team already using suitable Macs, try local document workflows in the Mac app on Pro before upgrading for cloud capacity, then compare actual credit use. The Mac pages do not specify a blanket credit exemption, so do not budget every Mac task as free.

Perplexity publishes these typical credit-billed task ranges:

• A quick fact or definition uses 15 to 70 credits, or $0.15 to $0.70.
• A multi-step research report uses 350 to 950 credits, or $3.50 to $9.50.
• A data dashboard uses 2,400 to 9,800 credits, or $24 to $98.
• An entire app or multi-day build uses 15,000+ credits, or $150+.

These are examples, not guaranteed task prices.

Those ranges turn "use the agent" into a budgeting problem. A local-services owner asking Computer for a quick definition may spend less than a dollar. A founder asking it to collect competitor data and create a dashboard can spend close to $100 on one run. The outcome, not the prompt length, drives cost.

Pro is especially easy to misread. It still has no recurring monthly credit allocation, but paid consumer Pro now includes weekly Computer usage before credits are spent. The rollout is gradual: check for Included in Pro on the Usage page described in the credit guide. This allowance resets weekly, does not roll over, and is excluded from promotional or non-paying Pro, Max, and Enterprise plans. With no credit balance, included usage runs at Light effort; with credits, other effort settings are available and consume included usage faster. The separate 4,000-credit onboarding bonus is worth $40 and expires in about 30 days. Neither the bonus nor an unspecified weekly allowance is a fixed monthly automation budget.

Max includes 10,000 monthly credits, but unused monthly credits do not roll over. A quiet month cannot fund a heavy quarter. Purchased credits are consumed after bonus and monthly credits and expire after one year of inactivity.

Purchased credits are not forfeited when you cancel. You need an active paid plan to use them, and they expire after one year without using Computer. Monthly plan credits end with the subscription. Do not treat a retained purchased balance as permission to keep running Computer on Free.

Running out creates an operational cost too. Credit-billed tasks pause, new credit-billed tasks are blocked, and work resumes once credits return. Eligible Pro users can also wait for their included weekly usage to reset. Auto-refill is off by default. If enabled, it can trigger at 500 credits on Pro or 2,500 on Max. Max starts with a $200 monthly cap on additional credit charges, adjustable up to $5,000. The subscription price is therefore not the maximum possible monthly bill.

For recurring work, use a three-part control:

1. Price one representative task from the account's usage detail, checking whether it ran locally or used cloud compute and whether included Pro usage covered it.
2. Multiply by the real monthly frequency, including retries and failed deliverables.
3. Set the auto-refill cap below the point where a human should reassess the workflow.

A B2B operator running ten credit-billed research reports a month faces $35 to $95 in example credit costs. A studio running four credit-billed dashboards faces $96 to $392 before the base subscription. These examples apply to work billed to credits after any included usage, not automatically to every task. Those examples explain why Computer can be valuable without being cheap.

Enterprise pricing buys governance seat by seat

Perplexity Enterprise Pro costs $40 per purchased seat each month or $400 per seat each year. Enterprise Max costs $325 per purchased seat each month or $3,250 per seat each year. The plans add organization controls, shared knowledge, admin-managed billing, and data that is not used for model training.

Annual Enterprise Pro works out to $33.33 per seat-month. Enterprise Max works out to $270.83. For ten seats, Enterprise Pro costs $4,000 annually versus $4,800 across monthly payments, saving $800. Ten Enterprise Max seats cost $32,500 annually versus $39,000 monthly over a year, saving $6,500.

Do not put every employee on Max by reflex. Enterprise organizations can mix Pro and Max seats. A research-intensive strategy group can use Max while occasional users remain on Pro. That design keeps the premium attached to people who can consume the higher limits.

The published capacity difference is substantial. Enterprise Pro includes 400 Pro Searches a week and 50 Research plus 50 Create-files-and-apps queries a month. Enterprise Max raises those figures to 4,000 Pro Searches a week and 500 monthly Research plus 500 creation queries. Those published counts are more useful for capacity planning than the vague consumer bands. Enterprise Pro includes 500 monthly Computer credits per seat; Enterprise Max includes 15,000.

The billing boundary has two wrinkles. Removing a member frees a purchased seat without reducing the bill or creating a prorated credit. To pay for fewer seats, an admin reduces the seat count in Billing; the reduction takes effect at the next monthly or annual renewal. Enterprise plans also include no API access or credits. Product API consumption receives a separate bill even when the organization pays for hundreds of app seats.

Eligible educational institutions and nonprofits can apply for Enterprise Pro at $30 per seat monthly or $300 annually. Students cannot claim that institutional rate directly; the individual Education Pro route is separate. Perplexity does not offer self-serve Enterprise Pro or Enterprise Max trials, so a team should discuss evaluation with sales and assess governance before committing seats.

API pricing is separate from every app plan

Perplexity API pricing is pay-as-you-go and completely separate from Free, Pro, Education Pro, Max, Enterprise Pro, and Enterprise Max. A $200 Max subscription does not fund product API calls. A developer must purchase API credits or configure billing independently.

The simplest API is Search API. Standard web search costs $5 per 1,000 successful requests, or $0.005 per request, with no token charge. Fast Search costs $1 per 1,000 requests. One successful request can contain up to five queries and still count as one billing unit. Fully packed, standard web search costs $0.001 per query.

There is a billing edge: a successful response is charged even when it returns no results. Invalid requests, rate-limited requests, and upstream failures are not billed. A builder should deduplicate and batch queries before sending them, then treat an empty successful result as paid work rather than as an error refund.

Search API returns raw results. It does not synthesize an answer. If the product needs model reasoning and tools, Agent API is the relevant comparison.

Agent API bills third-party model tokens at each model's published rates, then adds tool charges. A standard web search costs $0.0025, a URL fetch $0.0005, a people or finance search $0.005, and a sandbox session $0.03. Perplexity's current example uses GPT-5.6 Terra with 500 input tokens, 200 output tokens, and one standard web search for a total of $0.0059. One thousand similar runs cost $5.90, and $20 funds about 3,390 of them before usage changes.

Sonar Chat Completions support ended on September 27, 2026. Perplexity is gradually reformulating synchronous and streaming Sonar requests as Agent API requests; asynchronous Sonar requests are no longer supported. Use the Agent API migration guide and current API rates when budgeting an existing integration.

These numbers do not make the API a cheaper version of the app. The app bundles a user interface, Projects, files, Research, model routing, and consumer limits. The API supplies programmable units that a product team must orchestrate, monitor, secure, and support. Compare API spend to the cost of the application being built, not to the number of chats in a Pro account.

Perplexity versus ChatGPT, Claude, and Google AI on price

Perplexity Pro is priced almost exactly with its three serious individual alternatives. The deciding difference is the work surface, not the sticker: Perplexity Pro is $20 monthly, ChatGPT Plus is $20, Claude Pro is $20, and Google AI Pro is $19.99.

ChatGPT: choose it for the broad general work surface

ChatGPT Plus costs $20 a month and has no annual billing option. It adds broader model and tool access, higher limits, advanced reasoning, voice, images, file analysis, deep research, and custom GPTs. ChatGPT Free now includes unlimited GPT-6 text chats, so Plus earns its fee through tools and higher capability rather than ordinary conversation.

Choose ChatGPT Plus when the week mixes writing, analysis, voice, image work, files, custom assistants, and coding support. Choose Perplexity Pro when current source discovery and visible citations are the center of the job. For the full OpenAI price boundary, the ChatGPT Free versus Plus comparison is the better companion.

The price difference appears only on annual commitment. Twelve months of ChatGPT Plus costs $240 because OpenAI does not offer annual Plus billing. Perplexity Pro costs $200 upfront, saving $40. That saving matters only if Perplexity is the right workflow.

Claude: choose it for native Claude work

Claude Pro costs $20 monthly or $200 paid upfront annually, with the annual page displaying a rounded $17 monthly equivalent. It includes more usage, Claude Code, Claude Cowork, Claude Design, Claude Science, Projects, Research, and broader model access.

Choose Claude Pro when the work lives inside Claude's native coding, document, research, or project surfaces. Selecting a Claude model inside Perplexity does not reproduce Claude Code or Cowork. Choose Perplexity when the source trail and multi-model research wrapper matter more than the native Claude environment.

The price is a tie on both billing cadences. The decision flips on workflow, not cost. A senior builder who spends the week in Claude Code should not buy Perplexity as a substitute. A strategy operator who must expose current sources should not buy Claude merely because its model appears in Perplexity too.

Google AI Pro: choose it when the bundle replaces other spend

Google AI Pro costs $19.99 a month and includes 5 TB of storage, expanded access to Gemini 3.1 Pro, and AI features across Google products. That storage bundle can make its effective AI cost much lower for a household or operator who already needs the capacity.

Choose Google AI Pro when Gmail, Docs, Google storage, and the wider Google ecosystem are the operating environment. Choose Perplexity when cited cross-web research and preferred-model choice are the purchase. A $0.01 monthly sticker difference should never decide between them.

The normalized decision is simple:

• Sources first: Perplexity Pro at $20.
• General mixed work: ChatGPT Plus at $20.
• Native Claude coding and documents: Claude Pro at $20.
• Google apps plus storage: Google AI Pro at $19.99.

Perplexity is not the cheapest monthly product, nor does it need to be. It is the right $20 purchase when the output must carry a current source trail. The broader Perplexity review covers how that workflow behaves beyond price.

Annual billing saves two months but tightens the exit

Perplexity annual plans save two monthly payments, but the refund window is too short to treat annual checkout as a trial. Pro saves $40, Max saves $400, Enterprise Pro saves $80 per seat, and Enterprise Max saves $650 per seat.

The effective annual seat-month rates are $16.67 for Pro, $166.67 for Max, $33.33 for Enterprise Pro, and $270.83 for Enterprise Max. Annual billing is financially rational when the workflow and plan are already proven. It is a poor discovery mechanism.

Refund rights depend on location. Customers in the EU, UK, or Turkey can seek a refund within 14 days of monthly or annual purchase. Customers in South Korea or Brazil have 7 days. For other customers, Perplexity's general refund page says 24 hours for monthly purchases and 72 hours for annual purchases.

The Max page conflicts with that general policy by stating 48 hours for annual refunds outside the listed regions. Assume the shorter 48-hour Max window, save the checkout confirmation, and submit any request immediately. A buyer should not rely on the more generous wording after the charge.

Subscriptions renew at the chosen monthly or annual cadence unless canceled. Perplexity's billing guide sets out the cancellation and payment-channel rules. Perplexity does not support pausing. Cancellation returns the account to Free at the end of the paid cycle, so access continues until then. Deleting the account does not cancel the subscription.

Billing channel matters. Perplexity warns that updating payment or upgrading through an app store after starting on the web can create a separate subscription and another charge. Manage and upgrade through the original purchase channel unless the account page explicitly routes the change.

Promotional trials also require a payment method and auto-renew at the selected standard rate unless canceled before the trial ends. A free promotional period is therefore a deferred subscription decision, not a no-commitment account.

1. 1

   Prove the plan monthly

   Use one monthly cycle to measure blocked searches, Research usage, Computer credits, and deliverable value. Do not annualize feature curiosity.

2. 2

   Capture the checkout terms

   Save the annual total, tax, renewal date, refund wording, and purchase channel before confirming payment.

3. 3

   Cancel in the same channel

   Web purchases are managed on the Perplexity account page. App-store purchases are managed through that store. Confirm the renewal status after cancellation.

Who should choose each Perplexity plan

Choose Standard Free when basic cited search is the job and 3 Pro Searches a day plus 1 Research query a month cover the occasional deeper question. It is not a trial that must become paid.

Choose Education Pro when SheerID verifies student or educator status and the Pro workflow belongs to recurring study, teaching, or research. At $10 monthly, it is the strongest individual price.

Choose Pro when source-first research recurs every week, advanced-model choice or files matter, and saving 12 to 24 minutes a month covers the fee at the value of your time. It is the default paid recommendation. On a supported Mac, test local Computer workflows before paying for Max solely to expand agent capacity.

Choose Max when measured use exceeds Pro's advanced limits and Max-only access repays the $80 monthly-billing premium over Pro plus equal credits. A first-month plan that uses the expiring bonus can justify Max temporarily; compare actual credit spend after any included Pro usage.

Choose Enterprise Pro when a company needs centralized seats, shared internal knowledge, dedicated support, and enterprise data that is not used for model training. Governance is the product boundary.

Choose Enterprise Max only for the subset of organization users who need 4,000 Pro Searches a week, 500 Research queries a month, 500 creation queries a month, and the highest access. Mix seats instead of paying $325 for every user.

Choose the API when software, not a person, must call Perplexity. App subscriptions and API billing never substitute for each other.

Frequently asked questions

Is Perplexity free?

Yes. Perplexity Standard costs $0 and includes practically unlimited basic searches, 3 Pro Searches a day, 1 Research query a month, and limited basic file uploads. Stay free if those limits do not interrupt recurring work.

Is Perplexity cheaper than ChatGPT?

Their main individual paid plans both cost $20 a month. Perplexity Pro also offers $200 annual billing, while ChatGPT Plus has no annual option, so twelve months costs $200 versus $240. ChatGPT Free now offers unlimited GPT-6 text chats, while Perplexity Free is centered on practically unlimited basic cited search.

Is Perplexity really worth it?

Perplexity Pro is worth $20 when current sources, citations, files, and Research are part of a weekly paid workflow. Standard is better for occasional use. Max is worth $200 only when higher limits and exclusive access repay $80 beyond Pro plus the same Computer credits.

Why is Perplexity so expensive?

The $20 Pro price matches ChatGPT Plus and Claude Pro. The expensive tier is Max at $200, which buys the highest limits, priority access, Brain, and 10,000 monthly Computer credits. Those credits are worth $100, so the rest of the Max product must justify the premium.

Does Perplexity offer a student discount?

Yes. Education Pro costs $10 a month for students and educators verified through SheerID. It includes Pro plus Learn Mode and education-specific guidance. No public annual Education Pro price is listed.

Can I get a refund from Perplexity?

The window depends on region and billing cadence. EU, UK, and Turkey customers have 14 days; South Korea and Brazil have 7 days. Elsewhere, the general page says 24 hours for monthly and 72 hours for annual purchases, while the Max page says 48 hours for annual refunds. For Max, act within 48 hours.

Does Perplexity Pro include API access?

No. Every app plan bills separately from the API. Standard Search API costs $5 per 1,000 successful requests, while Agent API adds model and tool charges according to usage.

Did Perplexity pricing change in 2026?

The core individual prices remain $20 a month for Pro and $200 for Max. The surrounding cost structure has changed: paid Pro is gaining weekly included Computer usage, Max includes 10,000 monthly credits, and supported Macs can run Computer locally. Measure the work left for paid cloud capacity before upgrading.

What is the downside of Perplexity?

Individual limits are vague, Computer adds variable spend, API use is separate, monthly credits do not roll over, and purchased credits require an active paid plan even after cancellation. The live documentation still conflicts on the annual Max refund window. The product is strongest when cited research is the recurring job, not when one assistant must do everything.

Assign each AI subscription to a recurring business outcome before paying for overlap. Get the AI Tools Map for Business Owners.