Wall & Fifth

Consumer brands · Worldwide · November 2026

Language learning apps: who the AI engines name.

Duolingo leads this sector, named in 90% of 20 answers across ChatGPT, Gemini. 52 brands were named at least once.

Questions

10

Answers

20

Brands named

52

Engines

2ChatGPT · Gemini

The ranking · November 2026

52 brands, in the order the engines put them.

Score is 0–100. Mention share is the proportion of all answers that named the brand; first is how often it was the first name given. Per-engine columns show the rank within that engine alone.

#BrandVisibility scoreMentionedFirstAvg. positionChatGPTGeminiChange
1Duolingo
81
90%55%2.2#1#2–
2Babbel
76
90%30%2.5#3#1–
3Pimsleur
61
85%5%5.1#2#4–
4Busuu
55
80%0%4.7#4#3↑2
5LingoDeer
39
60%0%6.8#12#5↑13
6italki
37
60%0%7.5#7#6↓2
7Memrise
34
55%0%6.8#5#10↑3
8Preply
34
55%0%8.6#13#7↑3
9Rosetta Stone
31
45%0%4.8#8#8↓4
10LingQ
27
45%0%11.6#10#9↓3
11Mango Languages
20
30%5%10.3#6#29New
12Anki
19
30%0%8.8#11#11↑1
13HelloTalk
18
30%0%10.3#9#25↓4
14Tandem
15
25%0%11.4#14#34↑8
15Speak
12
20%0%11.3#23#12↓3
Show the remaining 37 brands
16Clozemaster
9
15%0%10.7#17#22New
17Drops
9
15%0%7.3#18#23New
18ELSA Speak
9
15%0%13.3#19#24New
19Mondly
9
15%0%7.3#22#30–
20Readlang
9
15%0%16.3#15–New
21Beelinguapp
6
10%0%14.0#29#20New
22Cambly
6
10%0%15.0#16–New
23Language Reactor
6
10%0%20.5#20–New
24Lingvist
6
10%0%9.5#21–New
25Speechling
6
10%0%9.5#24–New
26TalkPal
6
10%0%5.0–#13↓5
27Transparent Language
6
10%0%15.0#25–New
28TruAccent
6
10%0%5.5–#14New
29iOS
5
5%5%1.0#26–New
30Android
4
5%0%2.0#27–New
31Apple App Store
4
5%0%2.0–#15New
32Babbel Live
4
5%0%3.0–#16New
33Duolingo Max
4
5%0%3.0–#17↓18
34Google Play
4
5%0%3.0–#18New
35iPhone
4
5%0%3.0#28–New
36Super Duolingo
4
5%0%2.0–#19New
37Busuu Premium
3
5%0%8.0–#21New
38Easy Languages
3
5%0%18.0#30–New
39FluentU
3
5%0%23.0#31–New
40Kahoot!
3
5%0%10.0–#26New
41Language Transfer
3
5%0%21.0#32–↓24
42Lingoda
3
5%0%20.0#33–New
43LingoPie
3
5%0%7.0–#27New
44Loora
3
5%0%10.0–#28New
45PCMag
3
5%0%5.0–#31↓37
46Pearson
3
5%0%13.0–#32New
47Praktika
3
5%0%11.0–#33New
48Rocket Languages
3
5%0%9.0#34–↓28
49SpanishDict
3
5%0%25.0#35–New
50Transparent Language Online
3
5%0%23.0#36–New
51WordReference
3
5%0%24.0#37–New
52Yabla
3
5%0%22.0#38–New

Measured November 2026 · 20 answers · updated 1 November 2026 · change is rank movement against the previous period

What we asked

The 10 questions, put to every engine.

  1. What is the best language learning app to buy in the world this year?
  2. Which language learning app would you recommend for an adult beginner?
  3. What are the top language learning app brands in the world?
  4. What are the best alternatives to Duolingo?
  5. I'm an adult beginner in the world. Which language learning app should I get?
  6. Compare the leading language learning app brands.
  7. What is the best value language learning app in the world?
  8. Which language learning app has the best reviews?
  9. Language learning app: what are the main brands and how do they differ?
  10. Is Duolingo the best language learning app, or is there something better?

Where the engines looked

The sources cited most in this sector.

  • polychatapp.com4 cites
  • lingq.com4 cites
  • apps.apple.com3 cites
  • blog.duolingo.com3 cites
  • pimsleur.com2 cites
  • help.busuu.com2 cites
  • play.google.com2 cites
  • copycatcafe.com2 cites
  • youtube.com2 cites
  • babbel.com2 cites
  • memrise.com2 cites
  • memriseandroid.zendesk.com1 cites

These are the pages the engines leaned on when they answered. A brand that wants to be named here needs to be described clearly on them.

Reading the result

What this sector tells us.

Duolingo is dominant here: named in most answers and usually first. Babbel is close behind, which usually means the engines are reading the same comparison pages and splitting the difference.

52 brands were named across 20 answers, but the top 5 account for most of the recommendations. In AI search the long tail is not a tail; it is silence. A brand outside the top ten here is effectively absent from the buyer’s shortlist in this market.

The engines cited polychatapp.com more than any other source. For a brand trying to move up, that page and the next few on the list are where the work starts, together with a site that states plainly what the company does, for whom and where.

Ask Wall & Fifth about this sector

Questions

About this ranking.

Questions readers ask about the language learning apps measurement.

Read the full methodology
Which language learning apps do AI engines recommend most?

In November 2026, Duolingo was named in 90% of answers across Gemini, ChatGPT, with a visibility score of 81/100, ahead of Babbel (76) and Pimsleur (61).

How was this ranking produced?

We asked 10 unbranded buyer questions about language learning apps to 2 AI engines with live web search on, 20 answers in total, and recorded every brand each answer named in order. The score combines mention share, top-three share and first-mention share. The full method is on the methodology page.

Where do the engines get their recommendations from?

The most-cited sources in this sector were polychatapp.com, lingq.com, apps.apple.com. Brands that are described clearly and favourably on those pages are the ones the engines tend to name. That is usually the first place to look if a brand is missing from this list.

Do the engines agree with each other?

Not entirely. LingoDeer ranks 7 places apart between its best and worst engine this period. Each engine reads a different mix of sources and weighs them differently, so the per-engine columns matter as much as the overall rank.

My brand is missing or ranked lower than expected. What can I do?

Check the sources the engines cite in this sector and how your brand is described there, then look at how your own site states what you do, for whom and where. Wall & Fifth works with businesses on exactly this: rewriting the pages and entity signals the engines rely on, then re-measuring.

More sectors

Wall & Fifth

Want to be the name the engines give?

Wall & Fifth measures your brand against this sector's questions, finds the pages and signals the engines are missing, and rebuilds them. Then we run the measurement again.

Start with a measurement