Gemini vs Perplexity in 2026: Gemini Wins the Assistant Rules, Perplexity the Citation Rules

Disclosure: NOXI publishes this page. NOXI is a document maker, not an assistant, so it is not one of the two contenders and is not ranked here. It is scored once, after the citation verdict, as a reference row on the same seven citation rules: 53, with full marks for sources next to the claim, an export that keeps the links, and a stated behaviour when a fact cannot be confirmed. That is why the rules, the weights and the source of every cell are printed.

Two dated changes make this the week to compare them. On 26 October 2026 Perplexity said Deep Research and Model Council are rolling out as skills in Perplexity Computer, Pro and Max on the web first, with free usage on every plan and credits after it [12]. Three days later Google wrote that the Gemini app can now generate, preview and edit files right in the chat, for all users [7]. The pages that rank for this pair today were written between March and July 2026, before either change. Everything below comes from the two vendors' own pages, opened logged out on 5 November 2026, scored on the same seven assistant rules as our Perplexity vs ChatGPT and Claude vs Gemini pages, and on the seven citation rules of our ranking of AI tools with citations.

Gemini vs Perplexity at a glance (2026)

Gemini leads Perplexity by 29 points on the assistant rules, 94 to 65, and Perplexity leads Gemini by 8 points on the citation rules, 54 to 46: the same two products, two printed rubrics, two different winners.

ChatGPT and NOXI appear only after the citation verdict, as reference rows outside the head-to-head.

Who should pick what

The two verdicts point different ways, so pick by the job. Each line rests on one documented fact.

• If you want one assistant inside Google's apps, pick Gemini. It runs on the web, in Mac and Windows apps, on mobile and in Chrome, and Google AI Pro adds Gmail, Docs and the Google Antigravity coding platform: six of the six surfaces our rule counts [1][10].
• If you want a numbered link on every answer, pick Perplexity. Its help centre says "Each answer includes numbered citations linking to the original sources", and the Free card lists "Citations in every answer" [17][14].
• If you want the cheapest paid step, pick Google AI Plus, printed at "$4.99/ month" on Google's US page, with twice the usage of the free plan [1].
• If you want Deep Research without paying, start with Gemini. Google ticks Deep Research for users without a plan, with no published number and a warning that it may be unavailable during periods of high demand; Perplexity's Free plan lists Research queries at "1/month" [2][13].
• If you need peer-reviewed papers, pick neither first. Consensus, Elicit and Jenni AI share the top score, 93, in our citations ranking, where Perplexity has 54 and Gemini 46 [27].
• If the deliverable is a cited sheet or report you send as a PDF, use NOXI (reference row): every fact keeps its source on the page, and links stay clickable in PDF and Word [26].

Key numbers

The scores are this page's arithmetic: a higher score means more is documented, not better answers.

Perplexity vs Gemini Deep Research after 26 October

After 26 October both companies offer Deep Research on a free account, and both put a ceiling on it: Perplexity's Free plan lists 1 Research query a month and every plan spends free usage before credits, while Gemini lists Deep Research without a plan and publishes no number at all.

What Perplexity changed. Its article dated 26 October 2026 says Deep Research and Model Council are becoming built-in skills in Computer, Perplexity's agent that works with tools, and that neither feature is being removed [12]. The rollout starts with Pro and Max subscribers on the web; the desktop and mobile apps follow. Inside Computer, Deep Research can browse, use connectors and run applications to create files, and you can keep going in the same session, for example turning the report into a slide deck. Each prompt spends free usage first; credits start once that runs out, or when you switch model or effort level in a follow-up [12].

What did not change. The plans article, modified 29 October, still lists Free at 3 Pro Searches a day and 1 Research query a month, with image generation and file creation marked "No" [13]. Computer itself is on paid plans; some free accounts get a one-time allowance of Computer turns that does not reset [24]. So every plan has a free Deep Research allowance and then credits, which is not the same as Deep Research being free on every plan. The article prints "Last modified on October 26, 2026", and its UTC timestamp agrees.

Google's side. Gemini's limits page ticks Deep Research for users without an AI plan, on standard limits that refresh every 5 hours until a weekly limit, and warns that it may be unavailable during periods of high demand [2]. It searches Google by default and exports with "Export to Docs" [3]. Google publishes no number for free Deep Research.

Gemini's 29 October change: edit the file in the chat

Google's post of 29 October 2026 says you can now generate, preview and edit files right in the chat: expand to full screen and type, highlight text to ask about it, or ask Gemini to refine the file [7]. PDFs work on web and mobile; Docs, Sheets, slide decks, Markdown, TXT and code on the web only. Google's own example is a PDF study guide made from uploaded handwritten lecture notes, and the post says it reaches all users in every language where the app is available. File creation itself opened to all Gemini app users on 29 April 2026 [6]. It is one of the five capabilities our free-plan rule counts, while Perplexity's plans table marks "File and App Creation" as "No" on Free [13]. For slides from either assistant, see which AI can make a PowerPoint.

Both price pages now lead to credits

One quieter change runs through both ladders. Perplexity Pro lists 4,000 bonus credits and Max lists 35,000 bonus credits plus 10,000 monthly credits, spent in Computer [14]; the bonus credits are one-time and expire in 30 days [24]. Google's plans page says Gemini limits refresh every 5 hours until a weekly limit, and that you can extend them by purchasing AI credits [1]. On either side, the subscription price no longer tells you the whole ceiling. For the third assistant in this triangle, our ChatGPT vs Gemini page covers Google against OpenAI.

What changed recently

Five dated movements on the vendors' pages since 26 October, with the page's date and our read date.

Also true today: the newest entry in Gemini's release notes is dated 13 October 2026, with nothing later, and Perplexity's newest changelog entry is dated 19 October 2026 [11][23].

How we scored Gemini and Perplexity

Both are scored on documented capability from their own vendors' pages, opened logged out on 5 November 2026, on two printed rubrics, so a cell can only be filled by a sentence on an official page.

Half marks and ties. On rubric A every rating converts the same way: a 4 on a 20-point rule is 16 points, a 3 on a 15-point rule is 9. On an equal total the name decides, alphabetically; no tie arises at 94 and 65. Rubric K's levels and tie chain are printed with its table in the research section below. Neither rubric has a right-to-left rule, so no such reading applies here.

How the pages were read. In a real browser, in a fresh tab, logged out, between 15:41 and 15:54 UTC, with five pages opened again between 16:03 and 16:09 UTC; no cell moved. From the UAE, Google's default plans page served dirhams, so this page prints only the US English page's dollars. Perplexity served dollars. Nothing is converted.

Is Perplexity better than Gemini? On which rules

On the seven assistant rules, no: Gemini scores 94 and Perplexity 65, and the 29-point gap comes from five rules while the two are level on the other two.

The arithmetic. Gemini rates 5, 5, 5, 5, 4, 4, 5 on A1 to A7: 20 + 10 + 10 + 20 + 12 + 12 + 10 = 94. Perplexity rates 2, 3, 2, 5, 4, 3, 3: 8 + 6 + 4 + 20 + 12 + 9 + 6 = 65. The gap of 29 is A1 plus 12, A3 plus 6, A2 plus 4, A7 plus 4 and A6 plus 3; A4 and A5 are level.

Where Gemini gains its 29 points

A1 is the largest cell on the page, 12 points. Without paying, Gemini documents all five capabilities the rule counts. Perplexity's plans table marks image generation and file creation "No" on Free, and its pricing page leaves app connectors off the free column [13][14]; the pricing page's "Asset generation: Limited" yields to the plans table's specific "No", as the method says. A3 is 6 points because Google prints a 1 million token window for AI Pro and Perplexity prints none [2]: a documentation gap, not a measurement. A2 is 4 points: $4.99 a month against $17. A7 is 4 points: Gemini reaches six surfaces, including Gmail and Docs and Google's agentic development platform, while Perplexity reaches four [1][22]. A6 is 3 points: both offer an opt-out, and Google adds temporary chats that are kept out of training [9][15].

Where they are level: paid tools and add-ons

At the twenty-dollar step both plans list Deep Research, file creation, and image and video generation [15][13][1][2]. Both let individuals add a custom MCP server, each behind a fence: Google limits it to users 18 or over in the US, Perplexity to plans that include Computer [8][21]. If connecting outside tools is why you pay, compare those fences, not the totals.

Sensitivity readings, all printed

• Equal weights, each rule read as a fraction of its maximum: Gemini 0.943, Perplexity 0.629. The order holds.
• Without A1, one of the two heaviest rules: Gemini 74, Perplexity 57 of the remaining 80 points.
• Without A4, the other one: Gemini 74, Perplexity 45. Paid tools are where Perplexity is strongest.
• Single cells: Perplexity A7 at 4, if its enterprise Microsoft 365 and Google Drive connectors counted as an office integration, gives 67. The largest single move, Perplexity A1 from 2 to 5, gives 77. No single cell reverses the assistant verdict.

Perplexity's research strengths sit outside rubric A, which is why the next section scores the two on citations.

Gemini vs Perplexity for research: citations scored

On the seven citation rules Perplexity scores 54 and Gemini 46, and the whole 8-point gap is one rule: Perplexity's Academic search mode earns the middle mark on scholarly papers, while Gemini searches Google by default.

We re-applied the rules of our best AI with citations ranking on 5 November 2026, cell for cell. They are plain points, and the weights add to 100.

Half marks and ties. The middle levels are 10 of 20, 8 of 15 (half of 15 rounded up, as printed) and 5 of 10. On an equal total the higher K1 goes first, then K7, then K4, then K2, then the name. Every cell below that is neither full nor zero says why.

The arithmetic. Perplexity 20 + 0 + 8 + 8 + 10 + 0 + 8 = 54. Gemini 20 + 0 + 0 + 8 + 10 + 0 + 8 = 46. The gap is K3 alone. On the six other rules the two products document the same thing at the same level: links next to sentences, no passage view, a capped free plan, an export with sources, no citation style and a general warning.

One cell has two readings. Perplexity's pricing page marks free web answers "Limited", while its plans article promises very high limits on basic searches for Free [14][13]. We scored the stated cap, 8, the same way Gemini's standard limits are scored. At 15, Perplexity would score 61.

Sensitivity readings on the citation rules

• Perplexity K4 at 15 (the plans article over the pricing page): Perplexity 61, Gemini 46. Perplexity stays ahead.
• Gemini K2 at 15 (a documented passage view) or Gemini K3 at 15 (a paper database): Gemini 61, Perplexity 54. The citation verdict flips on either single cell.
• Perplexity K3 at 0 (Academic not counted as an academic mode): 46 each; the tie chain runs level through K1, K7, K4 and K2, and the name puts Gemini first.
• Equal weights: Perplexity 0.51, Gemini 0.44. The order holds.

A verdict that one cell can reverse is a narrow one, and we print it that way. Both products also warn, in their own words, that you should check what they cite; our guide to AI hallucinations covers how.

Reference rows, after the verdict

Two rows sit outside the head-to-head on the same seven rules, unranked: ChatGPT, the assistant most people compare both against, and NOXI, the document maker that publishes this page.

The arithmetic. ChatGPT 20 + 0 + 0 + 15 + 10 + 0 + 8 = 53. NOXI 20 + 0 + 0 + 8 + 10 + 0 + 15 = 53. In our full twelve-row citations ranking the two are level and NOXI is listed first on the tie chain, on K7 [27].

What the NOXI row rests on. When the subject needs current facts, NOXI searches the live web first, and every fact keeps its source on the page as an inline link, a footnote or a sources list; links stay clickable in PDF and Word. When a fact cannot be confirmed, NOXI leaves a clear field for you to fill instead of guessing, the behaviour K7 asks for. K4 is the middle mark because the free plan covers five documents a month. NOXI researches the live web and keeps each source on the page; a paper database, a passage view and citation styles sit outside what its product facts describe, so K2, K3 and K6 take the low mark [26]. It is a yardstick here, not a winner.

Gemini vs Perplexity Pro: the paid plans side by side

At the first paid step Google is cheaper, $4.99 a month for Google AI Plus; at the twenty-dollar step Perplexity Pro costs $17 a month billed annually against $19.99 a month for Google AI Pro; above that both ladders climb past $100 a month.

Twelve months, on the vendors' own figures. Perplexity Pro is 12 × $17 = $204, billed annually. Google AI Pro is 12 × $19.99 = $239.88 on the monthly price, $35.88 more, with no annual commitment printed. Google AI Plus is 12 × $4.99 = $59.88. Two Perplexity pages disagree on Max: 12 × $167 = $2,004 on the pricing page against "$2000/annually" in the help centre, $4 apart; "$200/monthly" makes $2,400 a year. We print both.

The fine print. Google allows cancelling at any time, with no refunds for partial billing periods except where the law requires, and its AI plans are for ages 18 and over [1]. Perplexity's Pro FAQ says searches may be used to improve its AI by default, with an opt-out in Account Settings [15]. Outside the United States, open both pages in your own country before you buy. For a wider set of plans side by side, see Claude vs Gemini, and for Microsoft's tiers after its rename, our sister page Copilot vs Microsoft 365 Copilot.

Worked example: one researcher, two subscriptions, twelve months

For a fictional freelance researcher who already uses Gemini for free, the setup that covers her client work at the lowest cost is Perplexity Pro on annual billing, $204.00 a year, with Gemini kept on the free plan.

Step one, count the research runs. Two briefs a week for four weeks is 2 × 4 = 8 deep research runs a month. Perplexity's Free plan lists 1 Research query a month, so it covers 1 of her 8 [13]. Gemini without a plan lists Deep Research with no number and a warning that it may pause at peak demand [2]. For client deadlines that is a risk she cannot plan around.

Step two, check the sources she needs. Both products put links next to sentences, the same 20 points on K1 [17][4]. What separates them for market sizing is data: Perplexity Pro adds premium sources such as PitchBook and Statista, which Google's plans page does not list [14][1]. Her briefs quote market figures, so that line decides.

Step three, price the options.

The arithmetic, checked twice. Google AI Pro costs $239.88 minus $204.00 = $35.88 more a year than Perplexity Pro. Taking both Pro plans costs $443.88, which is $239.88 more than Perplexity Pro alone. Adding Google AI Plus to Perplexity Pro costs $59.88 more, for twice Gemini's free usage.

Decision. Perplexity Pro on annual billing, $204.00 a year, for Deep Research beyond one query a month and for the premium data; Perplexity prints no exact Pro allowance, so she checks her usage after a month. Gemini stays free for drafts she exports to Docs and for editing PDFs in the chat [7]. She skips Google AI Pro: its 1M-token window suits long reports, and her briefs are short. The one-page price sheet her clients get each month she builds in NOXI, as in the build-it section below. Review on 1 December 2026.

What these rules don't measure

Fourteen rules leave out much of what decides a subscription, and several gaps favour Perplexity, the lower assistant score.

• Answer quality and accuracy. We score what is documented, not whether an answer is right. A link next to a sentence earns K1 even if the page behind it says something else.
• Which models answer. Perplexity says it picks among several top models per query, naming ChatGPT, Google Gemini, Anthropic Claude and NVIDIA Nemotron, and Pro lets you choose one [14]. Not scored.
• Premium data. Perplexity Pro and Max add PitchBook, Statista and S&P Capital IQ for finance and market work [14]. For an analyst that can outweigh every rule above.
• The Google bundle. Google AI Plus includes 400 GB of storage and AI Pro 5 TB [1]. Rubric A counts surfaces, not storage.
• Scholarly depth. Neither is a paper database. Google's Gemini Notebook shows the full quoted text when you hover a citation and gives no answer that is not in your sources, which scores it 68 on rubric K [28][27].
• Agents. Perplexity Computer and Model Council, and Tasks in Gemini on Google AI Ultra in select countries, are scored on no rule [24][12][1].

What would change this ranking

Three changes would move the numbers, each checkable on a named page.

• 1. Google publishes a free Deep Research quota, or removes Deep Research from the no-plan column of support.google.com/gemini/answer/16275805. Gemini's A1 could fall to 4, for 90. We re-check on 1 December 2026.
• 2. Perplexity's pricing page drops "Limited" from Free on "Get answers from the web", which would lift K4 to 15 and Perplexity to 61 on the citation rules, or adds image generation or file creation to Free, which would lift A1 and Perplexity to 69 or more. Page: perplexity.ai/hub/pricing. We re-check on 1 December 2026.
• 3. Perplexity publishes a context window for Pro on perplexity.ai/pro or its plans article (A3 to 3 or 5: 67 or 71), or Google documents the exact passage behind a Gemini source (K2 to 15: Gemini 61, and the citation verdict flips). We re-check on 1 December 2026.

Build it: a cited price infographic in NOXI

If the comparison ends in one sheet for a client, a class or a newsletter, make it with the Infographic format of NOXI 2.0, one of 14 formats from one request. The chart comes from a table of numbers you give it, and the file leaves as a PDF. One flow, six steps:

Gemini vs Perplexity: FAQ

Gemini vs Perplexity, which is better overall?

On seven assistant rules checked on 5 November 2026, Gemini scores 94 and Perplexity 65: Gemini leads on the free plan, entry price, long documents, surfaces and training control. On seven citation rules Perplexity leads 54 to 46, because of its Academic search mode. Pick by job: everyday assistant, Gemini; cited web answers, Perplexity.

Is Perplexity better than Gemini for research?

For cited web answers, slightly: Perplexity scores 54 and Gemini 46 on our seven citation rules, and the whole gap is Perplexity's Academic search mode. Both attach sources to sentences, both export reports, and neither shows the exact passage. For peer-reviewed papers, Consensus, Elicit and Jenni AI share 93 in our citations ranking. Checked 5 November 2026.

Perplexity vs Gemini: which should I pay for?

If you live in Google's apps, Google AI Plus at $4.99 a month is the cheapest paid step, and Google AI Pro at $19.99 adds a 1M-token context window. If cited research is the job, Perplexity Pro costs $17 a month billed annually and adds premium data sources such as PitchBook and Statista. Prices checked 5 November 2026.

Is Perplexity Deep Research free now?

Partly. Perplexity's help centre said on 26 October 2026 that Deep Research is rolling out as a skill in Perplexity Computer, starting with Pro and Max on the web, and that all plans include free usage, after which credits are used. The Free plan's table lists Research queries at 1 a month.

Is Gemini Deep Research free?

Yes, with limits Google does not number. Google's feature table ticks Deep Research for users without an AI plan, on standard limits that refresh every 5 hours until a weekly limit, and warns it may be unavailable during periods of high demand. Reports use Google Search by default and export to Google Docs. Checked 5 November 2026.

Gemini or Perplexity for a student?

Start with Gemini's free plan: Deep Research without a plan, file creation, and since 29 October editing files in the chat; Google's example is a PDF study guide from handwritten notes. Perplexity Education Pro costs $10 a month with SheerID verification. Google's US plans page sets its AI plans at 18 and over. For papers, see Consensus or Elicit.

Does Gemini cite sources?

Yes. Google's help page puts the Sources button "at the bottom of the response or in-line throughout the response". When Gemini quotes a long passage, it adds a link to that page in the sources list instead of highlighting the passage. So it earns full marks for placement and none for showing the exact passage. Checked 5 November 2026.

Which is cheaper, Gemini or Perplexity?

Gemini, at the first paid step: Google AI Plus is $4.99 a month on Google's US page, while Perplexity's cheapest general plan, Pro, is $17 a month billed annually. At the twenty-dollar step Perplexity is cheaper: $204 a year for Pro against $239.88 for twelve months of Google AI Pro at $19.99. Checked 5 November 2026.

What is Comet?

Comet is Perplexity's own web browser. Its page says it is available for Mac, Windows, iOS and Android, and its FAQ says Perplexity will always provide a free version that includes Comet. On our rules it counts as one of the six places an assistant works, alongside the web, desktop and mobile apps.

How much is Perplexity Pro vs Google AI Pro?

Perplexity Pro is $17 a month when billed annually, $204 a year, and its pricing page prints no monthly-billing price. Google AI Pro is $19.99 a month on Google's US page, $239.88 over twelve months. Google's plan adds a published 1M-token context window; Perplexity's adds 4,000 one-time bonus credits and premium data sources. Checked 5 November 2026.

Sources & methodology

Every price, plan line and cell comes from Google's or Perplexity's own pages, opened logged out on 5 November 2026 (times below, UTC). Prices are as served, never converted. Maya Ortiz is fictional; every calculation was recomputed twice.

• [1] Region label "United States"; Free "$0/ month with a Google Account" with Deep Research, image generation and editing, Gemini Live, Canvas, Gemini Notebook and 15 GB of storage; Google AI Plus "$4.99/ month", "Get 2x higher usage access than Free", video generation, 400 GB; Google AI Pro "$19.99/ month", "Get 4x higher usage access than Free", Gemini in Gmail, Docs, Vids and Chrome, Google Antigravity, 5 TB; Google AI Ultra "Starting at: $99.99/ month", "5x higher usage limits vs. AI Pro", "$199.99 / month", "20x", first access to Tasks in Gemini in select countries; limits refresh every 5 hours until a weekly limit and extend with purchased AI credits; cancel any time, no refunds for partial billing periods; plans for ages 18 and over; US dollars; opened 15:45 UTC, re-read 16:07 UTC · Google, Gemini plans (US).
• [2] Feature table "Without an AI plan": Canvas, Connected apps, Slide generation, Deep Research and image generation ticked, video generation from AI Plus up; "Your limit refreshes every 5 hours until you reach your weekly limit."; context windows of 32K without a plan, 128K on AI Plus and "1 million tokens" on AI Pro and AI Ultra; features needing more compute, like Deep Research, "may be unavailable during periods of high demand"; no free Deep Research number; opened 15:44 to 15:47 UTC · Gemini Apps Help, limits and features.
• [3] "By default, Gemini includes Google Search as a source for your research."; limits for "Daily research requests"; "To export the report, click Export to Docs"; opened 15:45 UTC · Gemini Apps Help, Deep Research.
• [4] The Sources button "at the bottom of the response or in-line throughout the response"; a long direct quote adds a link to that webpage in the sources list; opened 15:46 UTC · Gemini Apps Help, view related sources.
• [5] "Gemini Apps support most file types." · upload files; "Below the response, click Share & export" then Export to Docs · export responses; "Gemini can hallucinate and present inaccurate information as factual." · Gemini Apps Help, accuracy; opened 15:46 UTC.
• [6] Gemini creates PDFs, Word and Excel files and Docs, Sheets and Slides; "now available to all Gemini app users globally"; dated Apr 29, 2026; opened 15:46 UTC · Google blog, generate files in Gemini.
• [7] "You can now generate, preview, and edit files right within your chat"; full screen editing, highlight to ask; PDF on web and mobile, Google Docs, Sheets, slide decks, Markdown, TXT and code on the web only; a PDF study guide from uploaded handwritten lecture notes; available to all users in all languages where the Gemini app is available; dated Oct 29, 2026; opened 15:46 UTC, re-read 16:03 UTC · Google blog, edit and preview documents directly in Gemini.
• [8] "add a custom app with its Model Context Protocol (MCP) server URL"; "Be 18 or over and in the US."; opened 15:46 UTC · Gemini Apps Help, custom apps.
• [9] "Temporary chats are not used to train Google's AI models."; an opt-out setting; "Last updated: October 26, 2026"; opened 15:45 UTC · Gemini Apps Privacy Hub.
• [10] "Use the Gemini app on Mac" and "Use the Gemini app for Windows"; opened 15:46 UTC · Gemini app on Mac and Gemini app for Windows.
• [11] Newest entry "2026.10.13", nothing later; opened 15:52 UTC · Gemini app release notes.
• [12] Deep Research and Model Council "are rolling out as skills in Perplexity Computer, starting with Pro and Max subscribers on the web"; "All plans include free usage of Deep Research."; Deep Research remains included on Pro; neither feature removed; desktop and mobile apps to follow; turn a report into a slide deck in the same session; the plus (+) menu; free usage first, then credits, or credits when switching model or effort level; Model Council compares several models' answers; "Last modified on October 26, 2026", timestamp 2026-10-26T21:01:58Z; opened 15:41 UTC, re-read 16:04 and 16:27 UTC · Perplexity Help Center, Deep Research and Model Council in Computer.
• [13] Free: "Basic file uploads (limited)"; table Pro Searches "3/day", Research Queries "1/month", Image Generation "No", File and App Creation "No"; very high limits on basic searches (paraphrased); Pro "Opt-out AI training available"; Education Pro "($10/mo with verification from SheerID)"; Enterprise Pro "(starts at $40/month or $400/year/seat)"; "Last modified on October 29, 2026"; opened 15:42 UTC · Perplexity Help Center, which plan is right for you.
• [14] Free "$0 /month", "Good for limited daily usage", "Citations in every answer"; Pro "$17 /month when billed annually", "4,000 bonus credits", "Deep research", "Create polished documents & apps"; Max "$167 /month when billed annually", "35,000 bonus credits", "10,000 monthly credits"; table "Get answers from the web" Free "Limited"; app connectors not on Free; "Asset generation" Free "Limited"; FAQ: limited weekly usage on Free, inline citations on every answer, premium data sources PitchBook, Statista and S&P Capital IQ on Pro and Max, models orchestrated per query including ChatGPT, Google Gemini, Anthropic Claude and NVIDIA Nemotron; US dollars; opened 15:42 UTC, re-read 16:05 UTC · Perplexity pricing.
• [15] Pro "$17 /month or equivalent, when billed annually"; Max "$167 /month or equivalent, when billed annually"; FAQ: searches may be used to improve the AI by default, "You can opt out anytime in Account Settings."; no context window printed; opened 15:44 UTC · Perplexity Pro.
• [16] "Perplexity Max costs $200/monthly or $2000/annually."; "Last modified on September 14, 2026"; opened 15:43 UTC · Perplexity Help Center, Perplexity Max.
• [17] "Each answer includes numbered citations linking to the original sources"; "Last modified on October 26, 2026"; opened 15:41 UTC · Perplexity Help Center, how does Perplexity work.
• [18] Pro Search can "narrow your search to specific content types (e.g., Academic, Finance, your own files)"; "it's crucial to validate information by referencing the sources linked in your answer"; modified October 26, 2026 · what is Pro Search; "Choose sources" lists Web, Org Files, Web + Org Files and None; modified September 14, 2026 · choose sources; opened 15:42 UTC.
• [19] "export the final report to PDF or document"; "Last modified on October 20, 2026"; opened 15:43 UTC · Perplexity Help Center, what is Research mode.
• [20] Government, Academic or Trusted labels on cited domains; "Select or hover over a source to see its label"; modified 6 October 2026; opened 15:54 UTC · Perplexity Help Center, source labels.
• [21] "Connectors: remote MCP servers listed in an mcp.json file"; plugins with skills need "a plan that includes Perplexity Computer"; "Last modified on November 2, 2026"; opened 15:53 UTC · Perplexity Help Center, create or upload a custom plugin.
• [22] Comet "Available for Mac, Windows, iOS, and Android"; FAQ: a free version of Perplexity that will include Comet · Perplexity Comet; "Run Personal Computer on any Mac through the Perplexity macOS app." · Perplexity Personal Computer; opened 15:44 and 15:54 UTC.
• [23] Newest entry dated "10/19/26", "Ultrafast Computer models, Decisions API, and connector controls"; nothing later; opened 15:44 UTC · Perplexity changelog.
• [24] Computer is "available on paid plans"; some free accounts get a limited number of free Computer turns, a one-time allowance that does not reset; available on web desktop, iOS and Android; bonus credits one-time, expiring in 30 days; research, documents, scheduled and multi-step tasks; "Last modified on November 2, 2026"; opened 16:09 UTC, re-read 16:28 UTC · Perplexity Help Center, what is Computer.
• [25] "Select a citation to open its source."; "Search results and citations can be incomplete, outdated, or incorrect."; "Web search is available on ChatGPT Free" · OpenAI Help Center, searching the web with ChatGPT; deep research reports download "in multiple formats, including Markdown, Word, and PDF" with citations or source links · deep research in ChatGPT; chatgpt.com/pricing served UAE dirhams only, so no ChatGPT price is printed; opened 15:47 to 15:48 UTC.
• [26] NOXI 2.0, published 3 November 2026: 14 formats from one request, Infographic described as one tall or square sheet that explains with numbers and charts; live-web research with each fact's source on the page, links clickable in PDF and Word, a clear field to fill when a fact cannot be confirmed; a design library of more than ninety styles; charts built from a table of numbers in seven types, redrawn when the data changes; edit by chat or by hand, version history; export to PDF, PowerPoint and Word compared with the pages before download; no watermark; free plan of five documents a month; no credit card to start; a free account to create. The launch article is linked in the build-it section above.
• [27] Rubric K, its seven rules, levels and tie chain, and the full twelve-row table (Consensus, Elicit and Jenni AI 93; Gemini Notebook 68; Scite 55; Perplexity 54; NOXI 53; ChatGPT 53; Gemini 46), from our ranking of AI tools with citations of 5 November 2026, linked near the top of this page.
• [28] "To get the full quoted text, hover over any citation." and "If the answer isn't in the source material, it won't provide a response."; opened 15:47 UTC · Gemini Notebook Help, citations and answers from your sources.