Does Duolingo Work? What the Evidence Says (2026)

Last verified: September 20, 2026. Study details come from the papers’ abstracts, Duolingo’s own reports and blog, and secondary accounts where noted. Official benchmarks come from the US Foreign Service Institute, Cambridge English and the Council of Europe. Trustpilot figures change daily.

Does Duolingo work? For building a foundation, yes, but the evidence is thinner than the marketing suggests. Studies show that learners improve, mostly at beginner and lower-intermediate levels, and Duolingo now offers courses through CEFR level B2 in nine languages. But several of the best-known studies were funded by Duolingo, most are small, and an independent 2023 review concluded that “8 years after the start of research on Duolingo, we still have very little conclusive evidence about its effectiveness.” Duolingo works best as a daily habit and a source of vocabulary and basic grammar. It is not a substitute for speaking with people and consuming real content.

Duolingo app spotlight: free daily habit-building with 40+ languages
Duolingo: free daily habit-building

Quick verdict

Duolingo
What it does wellBuilds a daily habit, vocabulary and basic grammar, at no cost to start
Less well provenSpeaking and writing ability, and progress beyond the beginner stage
Evidence qualityMixed: several key studies are Duolingo-funded, small or short
How far it says it goesCEFR B2 in nine courses (April 2026); other courses lower
US App Store rating4.7 from 5.5 million ratings
TrustpilotAbout 1.6 out of 5 from about 8,100 reviews

How far does Duolingo say it can take you?

On April 22, 2026, Duolingo announced that nine of its most popular courses now “teach advanced content.” Its post says almost anyone learning English, Spanish, French, German, Italian, Portuguese, Japanese, Korean or Chinese can now learn through B2 on the CEFR scale, or a Duolingo Score of 129. Previously, its courses went up to A2, a Duolingo Score of 59.

For reference, the Council of Europe describes the levels like this:

Duolingo describes a B2 learner as an “independent user” who can interact with other speakers, understand certain complex texts and express viewpoints on a wide range of topics. That is what the company says its courses teach. It is a claim about content, not a measurement of how well the average learner does.

What does the research say?

The table lists the studies most often cited about Duolingo, with who paid for them and what they can and cannot show.

StudyFunding and designWhat it foundLimits
Vesselinov and Grego, 2012Funded by Duolingo; Spanish learners recruited by web adLearners improved on a placement test after using the appPer Stephen Krashen’s critique, there was no control group and only about 90 of 156 starters finished
Loewen et al., 2019 (ReCALL)Case study of 9 learners of Turkish over one semesterImprovement on language measures; a moderate positive link between time spent and gainsVery small; learners also reported frustration with the materials
Rachels and Rockinson-Szapkiw, 2018 (CALL)Quasi-experiment, 12 weeks, grades 3 to 4, Spanish; no conflict of interest reportedNo significant difference in Spanish achievement compared with classroom teachingChildren, and not randomized
Kessler, Loewen and Gönülal, 2023 (CALL)59 learners of Turkish, 8 weeks, Babbel compared with DuolingoBoth groups improved, with no significant difference between the appsSmall sample, short duration
Jiang, Rollinson, Pajak and Plonsky, 2021 (Foreign Language Annals)Duolingo-run; 225 volunteers (135 Spanish, 90 French) tested once with an ACTFL examMedian about 112 hours of use; reading around Intermediate Low to Mid and listening around Novice High, comparable to a fourth semester of university studyNo pre-test, self-reported prior level, participants differ from typical learners, only reading and listening tested; the authors say it should not be overgeneralized
Plonsky and Sudina, 2023Funded by a Duolingo grant but carried out independently; 287 learners over six monthsGains were linked to app time for written but not oral skillsVolunteers; oral gains not shown
Kim et al., 2026 (SSLA)French, 16 weeks; classroom, Duolingo and both groups; funded by the Duolingo Efficacy Research ProgramSimilar gains in all three groups; the app was “comparably effective”Funded by Duolingo’s research program
Shortt et al., 2023 (CALL)Review of 35 studies, excluding Duolingo-funded ones“Very little conclusive evidence” about effectivenessA review of the literature, not new data

Two things stand out. First, several studies are small, short, or funded by Duolingo. Second, the ones with controls and no reported conflicts of interest mostly find that Duolingo produces gains similar to other beginner-level methods, not dramatically better. The evidence is thinnest for speaking and for progress beyond the early levels.

How many hours does each level take?

Cambridge English publishes approximate cumulative guided learning hours for each CEFR level, as a guideline that varies with background, intensity and outside exposure. Duolingo’s own 2024 report says about 200 hours to reach the end of A2, on average over about 276 days. Putting that against a realistic study routine:

LevelCambridge guided hoursAt 30 minutes a dayAt 1 hour a day
A2180 to 200About 12 to 13 monthsAbout 6 to 7 months
B1350 to 400About 23 to 26 monthsAbout 11 to 13 months
B2500 to 600About 33 to 39 monthsAbout 16 to 20 months

These are rough estimates. App time is not the same as classroom guided hours, and progress depends on the language. The US Foreign Service Institute, which trains diplomats, estimates that its students reach level 3 on the government’s ILR proficiency scale in 24 to 30 weeks (552 to 690 class hours) for Category I languages such as Spanish, French and Italian, about 36 weeks (828 hours) for German, and about 88 weeks (around 2,200 class hours) for Category IV languages such as Japanese, Korean, Mandarin and Arabic. FSI students also do about 17 hours of self-study a week. So a B2 badge in Japanese or Chinese is a far bigger commitment than in Spanish.

What Duolingo does well

Where Duolingo falls short

Ratings: App Store vs Trustpilot

The two scores measure different things: in-app ratings come from active users, while Trustpilot attracts people with a billing or support problem. Read the Trustpilot number as a warning about billing and support, not a verdict on the lessons.

How to get more from Duolingo

  1. Use the free Practice Hub to review mistakes, practice speaking and listen.
  2. Add real conversation. A tutor, a language exchange or an AI speaking app fills the biggest gap. See Preply, Lingoda and our best language learning apps guide.
  3. Consume real content. Shows, podcasts and books build the listening and reading that lessons cannot.
  4. Review vocabulary outside the app. Free tools such as Anki help; see our best free language learning apps.
  5. Track hours, not streaks. A streak proves you showed up, not how much you learned.

Frequently Asked Questions

Does Duolingo actually work?

Duolingo works for building a foundation: studies show learners improve, and courses now run to B2 in nine languages. But several key studies were funded by Duolingo, most are small, and an independent review found very little conclusive evidence. It works best as a habit-builder alongside speaking practice.

Can you become fluent with Duolingo?

Not on its own. Duolingo says nine courses now teach through B2, the level of an independent user, but fluency also needs lots of listening to real speech and speaking with people. Duolingo’s own studies tested reading and listening, not open-ended conversation.

How long does it take to learn a language with Duolingo?

Duolingo’s own 2024 report says about 200 hours to reach the end of A2, roughly 276 days on average. Cambridge English guidelines put B2 at 500 to 600 guided hours, which is roughly 33 to 39 months at 30 minutes a day. Harder languages take longer.

Is Duolingo as effective as classes?

Studies point to similar gains rather than a clear winner. A 2018 study of children found no significant difference between Duolingo and classroom Spanish, and a 2026 French study funded by Duolingo’s research program found similar gains. Most studies are small, so the evidence is limited.

Is Duolingo better than Babbel?

A small 2023 study of Turkish learners found both apps produced gains with no significant difference. Duolingo is free with 40+ languages, while Babbel is a paid course with 14. Which suits you depends on your language and budget. See our Babbel vs Duolingo comparison.

How far can Duolingo take you?

Duolingo says nine courses (English, Spanish, French, German, Italian, Portuguese, Japanese, Korean and Chinese) now reach CEFR B2, a Duolingo Score of 129, with earlier courses stopping at A2. Other courses were not part of that announcement, so check your language.

Is Duolingo worth paying for?

Most learners can start free. Super removes ads and the Energy limit for about $96 to $120 a year, and Max, while it is offered, adds AI Video Call and Roleplay for about $168 a year. Explain My Answer and the Practice Hub are already free, so pay only if the extras solve a problem you have.

Why is Duolingo rated so low on Trustpilot?

Trustpilot attracts people with problems. Reviewers complain of unexpected annual charges after a free trial, crashes and poor support, and some say the app was ineffective for them. The US App Store rating from active users is 4.7 out of 5 from 5.5 million ratings.

Sources and how this was researched

This article compares information from published research and from Duolingo as of September 20, 2026. It does not claim first-hand testing. Some study details come from abstracts or secondary accounts, and Trustpilot figures change daily. Main sources: