Mango vs Duolingo: check your library card first

I have to tell you two things before anything else on this page is worth reading.

The first is a disclosure. I build a language app, Verbamor, and it competes for the same money these two do. You'll find it named once near the bottom, with its price and its limits. Every other claim here traces to a source you can open yourself.

The second is that I have not personally studied a course on either app. What follows is built from Mango's and Duolingo's own published pages, their App Store listings and the published research, all read on 21 September 2026 and quoted with that date. Three of the pages ranking above this one imply hands-on testing; one of them publishes no prices at all, which is a strange thing for a page about choosing where to spend money.

Do this before you read the rest

Mango Languages is free through thousands of American public libraries. Open mangolanguages.com/find-mango, type in your library, and you'll know in under a minute.

Mango's own page for librarians puts it plainly: "Give every patron free access to 70+ languages with their library card." The lookup page says "Thousands of public libraries, schools, colleges and other organizations give the people they serve full access to Mango, at no cost to them." Both read 21 September 2026.

Not a sample. Not a trial. The catalogue a paying subscriber gets, remote, from home, on the phone app, for the price of a card you probably already have.

I'm putting this first because it decides the question for a large share of readers, and because the competing pages bury it. One never mentions it, one gives it a passing clause, and one gestures at library access without a single figure attached. A comparison that walks you through $119.99 a year without telling you the same product may be sitting in your library catalogue has wasted your time.

If your library is on the list, stop reading, go sign in, and come back if you ever hit the ceiling described further down.

The short answer

Paying out of your own pocket, with no library route: start with Duolingo's free tier, and buy Mango only for a reason you can name.

Duolingo free is a complete course, ad-supported and energy-limited. It covers the same beginner ground Mango charges $119.99 a year for, and Duolingo's Spanish course now claims content through CEFR B2, which is further than Mango publishes any ceiling for. So Mango starts the argument owing you $119.99.

Three reasons make that a fair trade: you want the cultural and grammar notes Duolingo mostly skips, you want to record your own voice against a native speaker's, or you cannot stand streaks and leaderboards and will actually study without them. Those are real, and if one of them is you, Mango is a reasonable buy.

And the answer nobody selling you an app wants to give: if what you want is to hold a conversation this year, neither of these is the purchase that gets you there. A weekly hour with a tutor will do more than both subscriptions combined, and it costs more, and that's the honest shape of it.

What each one actually costs, in September 2026

Every comparison page on this query is quoting stale numbers, and I can be specific about how. Mango's structure changed. The pages ranking now describe a $11.99 single-language tier and a $19.99 all-languages tier, and then work out a yearly figure of $199.99 for all access.

There is no single-language tier on Mango's pricing page today, and there is no $199.99. Every plan includes all 70+ languages. Here is what the page says, read on 21 September 2026, with the App Store listing's term labels alongside it.

Mango's monthly and yearly figures come off its pricing page; the $119.99 and $149.99 yearly totals are the labelled Individual (Yearly) and Family (Yearly) in-app purchases on its US App Store listing. Mango's page also notes prices "may vary by country."

Now the Duolingo row, and why it looks evasive. Those five figures are the Super Duolingo in-app purchases on Duolingo's US App Store listing, and Apple prints no billing term against any of them. I won't tell you which is the monthly and which is the annual, because the store doesn't say, and the two pages that do tell you disagree with each other. One says Duolingo's annual works out to $7.99 a month; the other says $6.99. They cannot both be right, and neither shows its source.

Duolingo Max, the tier above Super, has no price on any page Duolingo owns. Every figure you've seen for it is secondhand. I'd rather leave a hole in the table than fill it with a number I can't stand behind.

One piece of arithmetic worth doing. Mango Family at $149.99 a year across 5 profiles is $30 per person per year, which is the cheapest paid route on this page by a distance, and it's the one nobody mentions because a household comparison doesn't fit a two-column template.

How they teach, and where they differ

Mango's method has a trademarked name, and its own how it works page describes the actual mechanism better than the name does: take real conversations, break them down piece by piece, then have you rebuild them in your own voice. Grammar explanations appear inside the lesson rather than as separate drills. Culture notes sit in the lesson too, on the theory that the meaning behind a phrase is part of the phrase.

Two Mango features are worth naming because Duolingo has no equivalent. Voice comparison records you and plays it back next to a native speaker, so you hear your own vowel against theirs. And Mango runs spaced repetition over its vocabulary, which is the right mechanism for keeping words rather than meeting them. (We've written up how modern spaced repetition schedulers work if you want the machinery.)

Duolingo's shape is the opposite. Short lessons, a streak, leaderboards, an energy meter that stops a free user who gets things wrong too fast. It's engineered for the habit, and it is extremely good at the habit, which is not a small thing given most people who buy a language app stop using it.

Duolingo also publishes more about its own machinery than almost anyone in the category. Its April 2026 announcement says nine of its courses now teach through CEFR B2, a Duolingo Score of 129. And Settles and Meeder published Duolingo's spaced repetition scheduler as a peer-reviewed ACL paper in 2016. You'll see it claimed that Duolingo has no real spaced repetition, and that's false.

Both companies advertise efficacy studies, and here the two of them are closer than either would like. Mango's 2019 study and Duolingo's 2012 study were run by the same two statisticians, Roumen Vesselinov and John Grego, and each was paid for by the app it evaluated. That's disclosed in both reports, on page 5, in nearly identical wording. I read all 16 reports from that team and wrote up what's in them. Treat "clinically proven" badges from either company accordingly.

What happens when you finish

Every page on this query concedes that neither app makes you fluent, and then none of them says what the ceiling feels like or why it's there. Here's both.

What you get: you read menus and signs, you follow written Spanish about a familiar topic, you recognise a few thousand words when you see them, and you understand slow clear speech aimed at a learner. That is genuinely worth having.

What you don't get: the word arriving in your mouth while somebody waits. You'll know you know it. It won't come. Under time pressure, with a real person's face in front of you, the gap between recognising a word and producing it is the whole gap.

The mechanism is in the exercise design. Both apps mostly ask you to choose: tap the matching tile, pick the right word, select the translation. Even the speaking exercises put the target sentence on screen before you say it, which means the retrieval was done for you and your mouth just followed along.

Kang, Gollan and Pashler tested this directly in 2013 with Hebrew vocabulary. One group heard each word and repeated it. The other had to produce it from memory before hearing it, with feedback. Two days later, on a spoken test, the group that had to produce the words beat the group that imitated them, with no loss of pronunciation quality. Their scope was picture-naming of single words, so it isn't proof about live conversation, and the direction is clear enough to plan around.

So what do you do at that point? Start producing language under mild pressure, which in practice means talking to somebody. A weekly tutor hour is the standard route and the one that works. If you want the range of what that costs, we compared the methods by the hour.

And keep one app running underneath it, because vocabulary decays without review whatever taught it to you.

Trials, cancelling and getting your money back

Nobody on this query covers the exit, so here it is.

Mango's trial. Both personal plans start with a free trial and Mango's FAQ says you can "Cancel anytime before the trial ends and you won't be charged." Mango does not publish how many days it runs, anywhere on its pricing page, so check the length on the signup screen before you commit. Its FAQ also says a plan change takes effect at the end of the billing cycle with no proration.

Duolingo's trial. Super Duolingo's US App Store listing carries a 14-day free trial.

Cancelling. Cancel wherever you bought. An App Store subscription cancels at Settings, your name, Subscriptions; a subscription bought on the company's website cancels in that company's account settings, and an App Store cancellation cannot reach it. Apple's own guidance is to cancel at least 24 hours before a free trial ends, which matters because both trials roll into a paid term automatically.

Refunds. Apple handles refund requests for App Store purchases at reportaproblem.apple.com, case by case, and neither company's published policy overrides that. I can't promise you a refund from either, and anyone who does is guessing.

Pick by your situation

Find the row that's you.

That last row is the one I built something for, so take the rest of this paragraph with the appropriate salt. Verbamor records your tutor lesson, pulls out the words and phrases you actually got wrong, and turns them into fill-in-the-blank cards with audio, scheduled by FSRS. It teaches you nothing new, which is the point: it holds what a tutor already taught you. Spanish, French, Italian, Portuguese and Russian, iOS only, US$19.99 a month or US$129.99 a year with a 7-day trial on the annual plan. Both are US prices, read off the app's own plan definitions on 21 September 2026; the App Store sets a different number in other countries. There's no free tier and no library route, so it's a worse deal than a library card by a wide margin, and it solves a different problem than either app on this page.

Questions people actually ask

---

Sources

Every load-bearing claim Verbamor makes is traced to its paper on the research page.