Russian vocabulary test with an adaptive word-count estimate

The test shows words of varying frequency and, from your answers, builds a statistical estimate using an IRT model (item response theory) — the same approach used in scientific vocabulary tests. The more rare words you know, the higher the estimate. The calculation runs in your browser.

Non-existent words that look real check your tendency to guess. If you mark made-up words as "known", the test understands that you might have guessed some of the "familiar" words too, and adjusts the result. This keeps the estimate honest.

This is a statistical estimate with a margin of about ±10%, not an exact count — measuring vocabulary "down to the word" is impossible in principle. But with honest answers, the estimate reflects the real size of your vocabulary well.

Estimates depend heavily on the method, so compare results only within a single test. An adult native speaker of Russian usually has tens of thousands of words in active and passive vocabulary combined; the more you read, the larger it gets.

Real words are taken from a frequency dictionary of the Russian language and the Russian Wiktionary (CC BY-SA license); the made-up ones are generated according to the rules of Russian spelling so that they look plausible but do not exist.