german5000
My words Take the test

How the list and the test work

Everything on german5000 hangs off one ranked list. This page says where the ranking comes from, who wrote the entries, how the test turns about thirty answers into a vocabulary size, and how review decides what you see next.

Where the ranking comes from

The ranks are counted, not chosen. We took two German corpora from the Leipzig Corpora Collection (Universität Leipzig): one million sentences of German news from 2023 and one million sentences of German web pages crawled in 2021. Each word form was counted per million words in each corpus, and the two rates were averaged, so neither corpus dominates. News alone over-ranks words like Prozent and Regierung; the web text pulls everyday words such as bitte and essen back up.

Upper and lower case were merged (Die at the start of a sentence counts as die). Numbers, punctuation and all-caps abbreviations were left out. Personal names, cities, companies and brands were then removed by hand, which is why the 5,000 kept words reach down to corpus rank 5,298. Countries, nationalities and languages were kept.

This is a list of word forms, not dictionary headwords. ist, war and sind are separate entries, each pointing back to sein, because a learner meets them separately. The list was last built on 2026-09-25.

What the list covers

Measured on the same two corpora, the top 100 forms make up 44.3% of all running words, the top 1,000 make up 64.4%, and all 5,000 make up 78.6%. The list has 2,202 nouns, 1,247 verb forms and 938 adjective forms. Written German is denser than conversation, so in everyday speech the same words cover more.

Who wrote the entries

The translations, parts of speech, CEFR levels and example sentences were written for german5000 with the help of an AI model (Claude, by Anthropic), checked by automated rules — every example must contain the word exactly as listed, every noun must carry its article — and are pending review by a native speaker. The example sentences are our own; none are copied from the corpora. CEFR levels are an editorial estimate of when a learner typically meets a word, not an official Goethe-Institut list. If you spot a mistake, it is ours.

How the test estimates your vocabulary

The test assumes the obvious: the more common a word is, the more likely you are to know it. It keeps a running estimate of your vocabulary size and picks each next word near the edge of that estimate, where your answer tells it the most. A right answer counts for less than it looks, because a quarter of blind guesses on four options are right, and “I don't know” counts as a clear no — so saying you don't know is always better than guessing. After about thirty words the range is usually narrow enough to stop. The result is the middle of that range. Your dashboard starts from that number: words above it go into your plan, and any word below it that the test caught you missing goes straight into review.

Levels on the result are a rough guide based on vocabulary size alone: under 500 words A1, under 1,200 A2, under 2,500 B1, under 4,000 B2, above that C1.

How review works

New words come from the top of your plan: the most frequent words you don't know yet. Each word you get right comes back after a growing gap — a day, then three days, a week, about two weeks, a month and longer. A word you miss comes back ten minutes later and starts the ladder again. A word counts as known once you have got it right three times in a row with at least three days between the last two. You can review with flashcards, multiple choice, typing or listening.

Analytics

We count page visits with Ahrefs Web Analytics. Each page loads a small script from analytics.ahrefs.com, which sends Ahrefs the page address, the referring page, your browser and device type and your approximate country. Ahrefs says the tool sets no cookies and does not identify individual visitors. Your saved words and test results are never included. Blocking that script does not affect anything on the site.

Audio

The play buttons use your own device's German text-to-speech voice. That works offline and sends nothing to us, but the quality depends on your device, and a few browsers have no German voice at all — the button tells you when that happens.

Credits

Lists built from this ranking: 100 most common, 500 most common, 1,000 most common, 5,000 most common, Most common verbs, Most common nouns, Most common adjectives, A1 vocabulary, A2 vocabulary, B1 vocabulary, B2 vocabulary, C1 vocabulary, Masculine nouns, Feminine nouns, Neuter nouns.