A native speaker costs 40% more in Korean and 34% less in Ukrainian

Across sixteen languages on one platform, the premium a native speaker commands runs from +40% to −34%, and the order tracks how concentrated that language’s native speakers are in expensive countries. Inside a single language it moves again, between countries where the language is natively spoken — a Dutch teacher is $40/hr in the Netherlands and $20/hr in Suriname.

Every figure below was read from the platform’s own public page on 24 August 2026. None of them are ours — we are not in these tables. Where we could not verify something first-hand it is named as unverified rather than filled in from hearsay.

Take the numbers and use them. No permission needed, no credit required, no link required. They came from public pages and you can re-read those pages yourself, which is the only reason a price table from a company that sells lessons is worth anything at all.

Download the whole thing as CSV — every row on this page, with its source and the date it was read, released CC0.

Preply publishes a native-versus-non-native comparison on every language’s tutor page. Read across all sixteen languages in our catalogue on a single day, it does not point the same way twice — and the order it falls into is not random.

If the number really tracks geography rather than nativeness, then it should still move inside a single language, between countries where that language is natively spoken. It does, and Preply publishes the breakdown itself.

Portuguese breaks the pattern. Mozambique tops that table at $32 — above Portugal, above the United States, above Brazil — and Mozambique is one of the lowest-income countries in the world. On a cost-of-living story it should be at the bottom.

We do not know why, and guessing would be worse than saying so. The most likely explanation is thin data: a country cell backed by a handful of tutors is an average over almost nothing, and a few specialists can carry it. That is a caveat for every table on this page, not just this one — Preply does not publish how many tutors sit behind each figure, so we cannot tell a robust average from a noisy one.

It is one row against twenty-six that point the same way, which is roughly what a real pattern with a noisy tail looks like. But it is in the table, at the top of its group, labelled — because a dataset that quietly drops its inconvenient row is not worth citing.

These pages are personalised by location. Read from a European address, Preply’s price filter came back in Danish kroner while its comparison table stayed in dollars. Someone reading from another country may be shown different numbers on the same URL, so date and locate any figure you quote.

Advertised is not paid. On Preply, lessons are bought as a subscription: its help centre states learners are charged every 28 days, that plan lessons do not carry over and expire at renewal, and that separately bought extra lessons survive two full cycles. A 365-day year holds just over thirteen 28-day cycles — that arithmetic is ours, not a claim they make. Cost per lesson actually taken is therefore reliably higher than any rate in these tables.

One day is one day. This is a snapshot, not a time series. Re-read the pages before citing the numbers as current.

KoloLingo is a marketplace for independent language teachers, so we sell the thing this page prices. Read it with that in mind.

The defence is that we are not in any table here, no figure is ours, every one of them is checkable on someone else’s site, and the numbers do not flatter us — the cheapest tutors in this dataset are cheaper than anything a new platform can conjure, and the incumbents’ selection is far larger than ours. If you want the argument that does cost us something, our fee comparison includes the row where a competitor beats us outright.

The per-language write-ups, with the reasoning rather than just the tables, are for Spanish, Norwegian and Ukrainian.

Corrections are welcome. If a figure here is wrong, or a platform publishes something we missed, write to hello@kololingo.com and we will change it and say who corrected us.