Perplexity is the citation-first answer engine — every claim links to a source, which is why it displaced Google for a lot of research work. Two things changed in 2026: the Comet browser dropped its paywall entirely and is now free on every platform, and Pro moved from a daily search cap to weekly usage limits, so heavy spikes hit the ceiling differently than before. Pro at $20 remains the tier that makes sense for most people; Max at $200 buys autonomous background agents, which is a different category of purchase.
Reviewed by EdGrows·Updated Aug 18, 2026Affiliate link — we may earn a commission
The Comet browser is free. It launched Max-only at $200 a month, which drew predictable criticism, and then dropped its paywall entirely during 2026. It now runs on Mac, Windows, iOS and Android, no account needed for basic browsing.
The nuance that most coverage skips: free gets you the browser and basic AI assistance. Browser Agent queries — the autonomous actions that make it more than a Chromium skin — are listed as unavailable on the free tier, with Pro and Max including access subject to limits. So it's a free browser with a paid autonomy layer, which is a fair deal and not quite what "the browser is free" implies on its own.
Pro moved from daily to weekly limits. The old fixed daily Pro Search cap is gone. In practice this helps steady daily users — a heavy afternoon no longer locks you out until tomorrow — and can catch out anyone whose work arrives in bursts, since a single intense week now meets a ceiling that daily caps would have spread across seven days.
One small practical note: the canonical pricing page is perplexity.ai/pro. The /pricing path has been returning a 404.
| Plan | Price | The reason to be here |
|---|---|---|
| Free | $0 | ~5 Pro Searches a day, cited answers, Comet browser |
| Education Pro | $10 | Same as Pro, student verification required |
| Pro | $20 ($200/yr) | Weekly Pro Search limits, model choice, Deep Research, Comet Plus included |
| Max | $200 ($2,000/yr) | Background Assistants, unlimited Labs, Perplexity Computer credits |
| Enterprise Pro | $40/seat | Governance on top of Pro |
| Enterprise Max | $325/seat | Heavy agentic and research capacity |
The gap between Pro and Max is the one that matters, and it's not a limits upgrade — it's a different product. Max buys autonomy: agents that run tasks while you're not watching, plus Perplexity Computer with a large credit pool.
For research, writing support and fact-checking — which is what the overwhelming majority of Perplexity users are doing — Pro covers it. Max is a purchase to make after repeatedly hitting Pro's walls, not in anticipation of doing so.
Using Perplexity as a daily research tool alongside ChatGPT and Gemini, the citation model is what makes it stick. Getting an answer with every claim linked to an openable source changes how much you can trust the output — not because the model is more accurate, but because verifying takes seconds instead of a separate search.
For checking current facts, comparing products, or getting oriented in an unfamiliar area quickly, it has consistently been the fastest route. The habit that formed without any deliberate decision: anything that needs to be right goes to Perplexity first, and anything that needs to be written goes elsewhere.
Deep Research is strong as a first pass. Ask a broad question and it comes back several minutes later with a structured, cited report that covers the ground properly. It works best as scaffolding for your own work rather than as a finished product — the coverage is genuinely good, and every so often a summarised claim reads a little further than the source supports, which is exactly why the inline links matter.
Model Council is a nice touch for contested questions: running the same query across several frontier models and seeing where they diverge is more informative than any single confident answer.
The Comet browser is pleasant and the sidecar assistant that reads the current page is useful in ordinary browsing, though it hasn't replaced a normal browser day-to-day.
The clearest limitation in practice is that Perplexity is a search product first. Asking it to write something substantial or work through code is possible and not what it's good at — the answers stay short, sourced and factual, which is the right design for its job and the wrong shape for drafting.
Anything requiring verification. Research, fact-checking, current events, competitive analysis, due diligence. The citations aren't a feature, they're the product.
Getting oriented fast. Unfamiliar topic, need the landscape in ten minutes with sources — this is the best tool for that specific job.
Academic work at $10. Education Pro is half price and citation-first output is exactly what papers need. Best-value AI subscription for students, comfortably.
Model choice in one place. Selectable frontier models plus Model Council for contested questions, without three subscriptions.
Writing and coding. It's a search engine with an AI interface, not an assistant. Claude for writing, ChatGPT for general work.
Workflow automation. Perplexity answers questions; it doesn't act inside your tools. It can't tag a ticket, update a CRM or close a conversation.
Budget-sensitive casual use. The free tier's ~5 Pro Searches a day is real but thin. DeepSeek is free without caps if sourcing isn't your priority.
Anyone considering Max casually. $200 a month for background agents is a specific purchase for a specific workload. Most people should not start there.
Against ChatGPT: ChatGPT is the general assistant with the wider ecosystem, and its search is capable. Perplexity is purpose-built around citations, and for verification work that focus wins. Same $20 price, complementary rather than competing — a lot of people pay for both without feeling silly about it.
Against Gemini: Gemini bundles Deep Research into a broader product with Workspace integration and video generation at the same price. Perplexity is sharper on sourcing and model choice. If you're already paying Google, Gemini's research covers a lot of the same ground.
Against You.com: You.com competed here directly and has since pivoted its focus toward developer APIs and enterprise research agents, which effectively leaves Perplexity as the consumer answer-engine default.
Against Felo AI: Felo organises search results as mind maps for visual topic exploration, which is a different job to Perplexity's direct sourced answers. Explorers may prefer Felo; verifiers want Perplexity.
Against Undermind: Undermind goes deeper on scientific literature specifically. For serious academic search, it's more thorough; Perplexity is broader and faster.
| Plan | Monthly | Annual | Key inclusions |
|---|---|---|---|
| Free | $0 | — | ~5 Pro Searches/day, citations, Comet browser |
| Education Pro | $10 | — | Full Pro features, student verification |
| Pro | $20 | $200 | Weekly Pro Search limits, frontier model choice, Deep Research, Comet Plus |
| Max | $200 | $2,000 | Background Assistants, unlimited Labs, Perplexity Computer credits |
| Enterprise Pro | $40/seat | — | Governance, admin controls |
| Enterprise Max | $325/seat | — | Highest agentic and research capacity |
| Comet Plus | ~$5 | — | Publisher content, included in Pro and Max |
| Sonar API | Per token | — | From ~$1/1M tokens, separate developer billing |
Checked August 2026. The Comet browser became free on all platforms during 2026, and Pro moved from a fixed daily Pro Search cap to weekly usage limits. Pricing trackers logged a lineup trim in early July, so verify any plan not listed here. The canonical page is perplexity.ai/pro — the /pricing path returns a 404. Sonar API billing is separate from consumer subscriptions.
Pro at $20 is the answer for almost everyone. Max is a different purchase for autonomous agent work, and jumping there before hitting Pro's limits is a tenfold overspend.
Students should take Education Pro. $10 for citation-first research is the best-value academic AI subscription available.
Install Comet regardless of your plan. It's free, and the sidecar assistant is useful even without Browser Agent access.
Annual saves about 17% on Pro — $200 against $240 — once you know you'll keep it.
Perplexity is the best sourced-answer engine available, and the reason is a design decision rather than a model advantage: every claim links to something you can open. That single choice makes it faster to trust than any general assistant, because verification is a click instead of a separate search. For research, fact-checking, competitive work and getting oriented in unfamiliar territory, it's the tool that changes how the work feels.
The 2026 changes mostly improve it. The Comet browser went free on every platform, removing the one objection that dominated its launch coverage — with the honest caveat that Browser Agent autonomy still requires a paid plan. Pro's shift from daily to weekly limits suits steady users and can pinch burst-heavy ones. And Education Pro at $10 makes it the obvious pick for anyone writing papers.
Two things to be clear-eyed about. Max at $200 is a different product, not a bigger Pro — it buys unattended background agents, and anyone not running that workload is overspending by a factor of ten. And Perplexity is a search product, so asking it to write or code produces short, sourced, factual output that's exactly wrong for drafting. That's a scope limitation rather than a flaw, but it means Perplexity usually sits alongside a general assistant rather than replacing one.
For research-heavy work, recommend at Pro without hesitation, or Education Pro if you qualify. For writing and coding, pair it with something else.
Note: Perplexity does not currently have an affiliate program with AIVario. We earn no commission, and this rating carries no commercial incentive.
Best for: Research and fact-checking, students and academics, competitive and market analysis, anyone who needs to see where an answer came from Not ideal for: Long-form writing, coding assistance, workflow automation inside other tools, casual users who'll hit the free tier's daily cap Bottom line: The citation-first answer engine remains the best of its kind, the browser is now free, and Pro at $20 is the tier — Max at $200 solves a problem most people don't have.
Free at $0, Pro at $20 a month or $200 a year, Max at $200 a month or $2,000 a year, Education Pro at $10 for verified students, and enterprise tiers from $40 per seat (Enterprise Pro) to $325 per seat (Enterprise Max). The Comet browser is free on all platforms, with Comet Plus at around $5 a month for premium publisher content — bundled free into Pro and Max. Note that the canonical pricing URL is perplexity.ai/pro; the /pricing path has returned a 404.
Yes, genuinely — it dropped its paywall in 2026 and runs on Mac, Windows, iOS and Android without an account for basic browsing. What isn't free is the autonomous part: Perplexity's plan table lists Browser Agent queries as unavailable on the free tier, with Pro and Max including access subject to limits. So the accurate description is a free browser with free basic AI assistance, where the feature that makes it meaningfully autonomous requires a paid plan.
Perplexity moved away from a fixed daily Pro Search cap to weekly usage limits. The practical difference is that a heavy research day no longer locks you out until tomorrow — but a heavy research week can hit a ceiling that a daily cap would have spread out. For steady daily use it's an improvement; for burst-heavy work it's worth watching. The free tier remains limited to roughly five Pro Searches a day.
Only for a specific kind of user. Max buys autonomous Background Assistants that run tasks unattended, unlimited Labs, Perplexity Computer with a large monthly credit pool, and the highest model access. That's a genuine capability jump, not just higher limits. But for research, writing and fact-checking — what most people actually use Perplexity for — Pro at $20 delivers the overwhelming majority of the value. Max is a purchase you make after you've hit Pro's walls repeatedly, not before.
Its own Sonar family plus selectable frontier models — GPT-5, Claude Opus and Gemini among them — on paid tiers. There's also Model Council, which runs a question across several frontier models simultaneously and synthesises where they agree and diverge. The catch worth knowing is that on default searches the interface doesn't always show which Sonar variant answered, so 'which model wrote this' isn't always visible.
Perplexity is built around citations; ChatGPT added search to a general assistant. For any question where you need to verify the answer — research, fact-checking, current events, competitive analysis — Perplexity's inline source links make checking fast, and that's the entire product philosophy. For writing, coding and open-ended conversation, ChatGPT is the better tool. Plenty of people pay for both and use them for different halves of their work.
A multi-step research mode that runs dozens of searches, reads across sources and returns a structured, cited report rather than an answer paragraph. It takes minutes rather than seconds. It's genuinely strong for a first pass on an unfamiliar topic — coverage and structure are good, and every claim traces back to something you can open. As with every tool of this kind, the citations need spot-checking, because a source being linked isn't the same as it saying what the summary claims.
Yes — Education Pro at $10 a month with student verification, half the standard Pro price. For anyone writing papers, that's arguably the best-value AI subscription available, since citation-first output is precisely what academic work needs. Verification is handled through a third-party service, so have your institutional credentials ready.
The fastest way to see what the literature says about a claim — provided you read the meter as a signal, not a verdict.
Still the best free AI assistant available, with no paid tier to upsell you. API pricing got more complicated in August.
Google's source-grounded research tool. The free tier limits how much you use, not what you can do.