German · vocabulary list · 4,214 notes · 8,428 cards · analyzed November 5, 2026
A large, frequency-ordered German vocabulary deck with audio, an image and an example sentence on every note; worth studying once you reorder it, because the default order opens with the hardest words.
Scored as a main course. This deck is a word list; if you use it next to a course, its sentence quality, grammar handling and context scores matter less.
80% of the score. These concern the learning material itself — the part that cannot be fixed without rewriting the deck.
| Criterion | Weight | Score | |
|---|---|---|---|
| Gradual progression and comprehensibility | 20% | 6/10 | |
| Sentence quality and usefulness | 20% | 8/10 | |
| Vocabulary selection and frequency | 15% | 6/10 | |
| Accuracy | 15% | 6/10 | |
| Grammar handling | 10% | 3/10 |
20% of the score. Worth having, and mostly addable later.
| Criterion | Weight | Score | |
|---|---|---|---|
| Context-based learning | 6% | 4/10 | |
| High-quality audiopresence only | 6% | 10/10 | |
| Card design and study usability | 5% | 6/10 | |
| Imagespresence only | 3% | 10/10 |
The deck is solid on coverage. It teaches 97% of the 1,000 most common German words and 94% of the top 2,000, and every note carries audio, a picture, an inflection field ("hat gesagt"-style principal parts such as "nimmt mit; nahm mit; hat mitgenommen") and a short, natural sentence like "Ich sorge für meine Schwester." Its weaknesses are sequencing and thin teaching. In the default order you meet Level160 words first (zerreißen, die Zielsetzung, zuführen, zurückweisen) and only reach ein, zu and ich after about 5% of the deck. Only 59% of the sentences add at most one new word, and the new-word rate stays near 16-17% through the second half. The sample shows a few slips: "das Nerv", "Ergeht" with a missing space, "occurence" and "errupt" in glosses. It teaches no grammar beyond a bare inflection field, and each sentence stands alone with no dialogue or story.
Use it if you already read some German, want a large vocabulary base ordered by frequency, and are willing to reorder the deck first. The tag path is clearly better at the start (54% of first-tenth words from the top 1,000, against 37% in the default order), so follow it. It suits intermediate learners who pair it with a grammar course or textbook, and it will not teach you how German works. Absolute beginners who want guided progression, learners who want grammar explained, and anyone who wants connected texts or dialogues should look elsewhere.
Frequency order is a sound principle and the new-word rate settles to a stable 16-17%, but in the default order the deck opens with about 5% of its hardest vocabulary (Level160-169) before the basics. Only 59% of the example sentences add at most one new word, so many examples bring in extra unknown words beyond the headword.
Eric hat sich seine Jacke beim Klettern zerrissen.card #1 — The first card a learner sees is a Level160 verb whose example sentence is entirely new.
Level001 6%, Level025 20%, Level049 34%, Level073 48%, Level097 63%, Level121 77%, Level145 91%, Level169 5%.from the measurements — The default order puts the highest levels first, so the basics (Level001) begin only after about 5% of the deck.
Im Kino ist es dunkel.card #3 — Even in the tag-ordered path, a bare function-word card like 'im' has an example that adds several new words, though the sentences are short and simple.
The new-word rate falls from 37.8% to 16.8% by part 5, then holds between 15.7% and 17.4% (median 16.6%) for the rest of the deck.from the measurements — The overall curve has the right shape, a drop followed by a stable low plateau.
1 | 421 | 37.8% | 410 | 37.8% | 24.1% | 7.1 | 1,133from the measurements — Part 1 has only 24.1% of sentences with at most one new word, much weaker than the later parts (54-73%).
The sentences are short, natural and built around the target word, and they sit on the card front where you actually read them. A few are stilted, carry typos, or are harder than the word they illustrate.
Nebenbei gesagt, Messer sind oft scharf.card #3310 — A stilted, contrived sentence with unnatural word order for the word nebenbei.
Ergeht, ohne etwas zu sagen.card #412 — A typo (missing space) in a taught sentence.
Das Gericht hob die einstweilige Verfügung gegen die Veranstalter der Demonstration auf.card #1056 — A legal-register sentence considerably denser than the noun it illustrates.
Er hat zwar Lust zu kommen, aber keine Zeit.card #437 — A natural, useful sentence that shows how the particle zwar works with aber.
Du lenkst schon wieder vom Thema ab.card #11 — A natural spoken-register sentence that models the split separable verb.
The deck covers common German vocabulary well (97% of the top 1,000 and 94% of the top 2,000 words), but the default order opens with about 200 notes of advanced Level160-169 words before the basic Level001 words like ein, zu and ich. It is a frequency-ordered course, but the opening misleads, so it lands in the middle band.
zerreißen ... Level: 160card #1 — The first card a learner meets is a Level160 verb, 'to tear (up, to pieces)', not a basic word.
German: der Absolvent ... English: graduate ... Level: 161card #14 — The opening stretch is specialist vocabulary (Absolvent, Abkommen, Auswertung, Bedrohung) from the highest levels.
The default order does not follow them: in 10% of the pairs of notes from different levels, the higher level comes firstfrom the measurements — The default order partly ignores the deck's own level progression, which is why the opening is out of place.
top 1,000: the deck teaches 973 of 1,000 (97%)from the measurements — Coverage of the most common words is strong, so the weakness is order, not content.
German: ein ... Level: 1from the measurements — The true basics (ein, zu, im, auf, werden, ich) sit in Level001, which the default order puts after the opening advanced block (path #1).
Most of the sample is accurate, including the conjugation strings and plurals. But a handful of real slips show up across about 90 notes, among them a wrong article on a noun, a missing word in a sentence, and typos in the sentences and glosses.
German: das Nervcard #3613 — Wrong grammatical gender on a headword; the correct article is 'der'.
Wasser muss man Energie zuführen, um es zum Kochen zu bringen.card #5 — The dative article is missing from the sentence ('Dem Wasser'), so the example is ungrammatical.
Ergeht, ohne etwas zu sagen.card #412 — Two words run together ('Er geht'), a typo in the target-language sentence.
Erwin und Brigitte gehen heute Abend in den Auerbachkeller essencard #18 — Sentence has no final punctuation, unlike its neighbors.
English: 1) occurence; 2) processcard #1869 — The English gloss misspells 'occurrence'; the gloss 'errupt' for ausbrechen has the same kind of error.
This is a frequency-ordered vocabulary list with one example sentence per word. No grammar syllabus is visible in the sequence, and the only grammar help is a morphology field of inflected forms with no explanations or tables.
Plural and inflected forms: zerreißt; zerriss; ist zerrissencard #1 — The one grammar aid is a bare list of verb forms, with no explanation of strong verbs or auxiliary choice.
Um die eigene Identität zu wahren, ist es wichtig, sich von anderen abzugrenzen.card #9 — Complex infinitive clauses and a reflexive separable verb appear with no prior structure-building.
Ich gehe davon aus, dass du mitkommst.card #802 — A dass-clause and a separable verb appear in a frequency-ordered slot, with no sequence leading up to them.
Sie hat sich im Spiegel gesehen.card #10 — Perfect tense and reflexive use appear on the tenth path note, introduced only as an example for "sich".
Most common tags: noun, Level160, Level161, Level162, Level163, Level164, Level165, Level166from the measurements — Tags carry only frequency level and part of speech, and no structure or grammar tags.
Each card pairs a headword with one standalone example sentence. Nothing links the sentences into a dialogue, story or source course, and the only grouping is frequency-based level tags with no thematic content.
Note types in use: Memrise - 4000 German Words by Frequency-7eb67-110c6: 4,214 notes, 2 card template(s)from the measurements — The deck is a single flat note type under a single deck, with no lesson or source-course structure.
Der Mann im grünen Jackett ist der Assistent vom Professor.card #20 — A self-contained example sentence beside unrelated neighbors ('die Arbeitskraft', 'ästhetisch'); there is no connected scene.
Im Sommer habe ich Geburtstag.card #1059 — Consecutive notes (sorgen, Fach, Verfügung, Menge, Verbindung, Sommer) are unrelated, so no topic or situation carries across cards.
Most common tags: noun, Level160, Level161, Level162, Level163, Level164, Level165, Level166, Level167, Level168, Level001from the measurements — The tags only mark frequency levels and parts of speech, not themes, situations or a source.
Sie sitzt ganz allein an der Bar und trinkt einen Schnaps nach dem anderen.card #25 — Some sentences carry a vivid context of their own, though it does not connect to any other card.
4,214 of 4,214 notes (100%) carry a sound file, 4,214 distinct recordings. This is presence, not quality: the tool reads the cards and cannot listen to them, so a studio recording by a native speaker and a decade-old robot voice score the same here. Play a few cards yourself before you trust it.
The two cards have clear directions and a sensible reveal, but the hard-coded "Plural:" label sits on verb conjugations, the headword is not marked in its sentence, and the styling is minimal.
{{#Plural and inflected forms}}<div class="plural">Plural: {{Plural and inflected forms}}</div>{{/Plural and inflected forms}}from the measurements — The label is fixed as "Plural:" for every note, including verbs whose field holds conjugated forms.
Plural and inflected forms: zerreißt; zerriss; ist zerrissencard #1 — A verb note whose field would be shown under a "Plural:" label.
{{English}}<br/> {{#Picture}}{{Picture}}<br/>{{/Picture}}from the measurements — The English -> German front is a clean cue of gloss plus picture, with the German answer withheld.
{{German}}{{#Audio}}<div style="display:none;">{{Audio}}</div>{{/Audio}}<br/>from the measurements — Audio is hidden with display:none, so there is no visible control to replay it.
Sample sentence: Der Beschuldigte weist die Vorwürfe zurück.card #6 — The target word is split and conjugated inside the sentence, with no marking of which words belong to it.
4,213 of 4,214 notes (100%) carry an image, 4,214 distinct files. Full marks at 50%, since images suit concrete vocabulary and add little to grammar or long sentences. Presence, not quality: the tool cannot see the pictures, so it does not know whether they illustrate the word or are decoration.
The article lists these as convenient but keeps them out of the weighted score.
The deck split into ten equal parts in study order. The new-word rate is the share of words in each part that had not appeared earlier; a deck that builds properly trends downward.
Format: 4,149 of 4,208 measured notes (99%) are sentences of 4 words or more; 59 (1%) are single words or short phrases. Across the whole deck, 59% of the sentences add at most one new word. 4,149 of the sentences are examples on a card that names the word it teaches; there the headword is the card's one new item, so an example passes only if it brings in no other unknown word.
part | notes | new-word rate | sentences | in sentences | <=1 new word | mean words | vocab so far -----+--------+---------------+-----------+--------------+--------------+------------+------------- 1 | 421 | 37.8% | 410 | 37.8% | 24.1% | 7.1 | 1,133 2 | 421 | 22.1% | 411 | 22.0% | 54.5% | 6.7 | 1,753 3 | 420 | 20.1% | 412 | 20.0% | 58.7% | 6.7 | 2,319 4 | 421 | 19.0% | 419 | 19.0% | 57.0% | 7.1 | 2,889 5 | 421 | 16.8% | 417 | 16.7% | 63.5% | 7.4 | 3,418 6 | 421 | 16.9% | 416 | 16.9% | 57.2% | 7.7 | 3,975 7 | 421 | 16.5% | 417 | 16.4% | 68.6% | 6.6 | 4,443 8 | 420 | 17.4% | 414 | 17.4% | 63.3% | 6.9 | 4,960 9 | 421 | 15.7% | 418 | 15.7% | 72.7% | 7.0 | 5,431 10 | 421 | 16.4% | 415 | 16.3% | 66.7% | 7.2 | 5,941 'sentences' counts the notes in that part holding at least 4 words; 'in sentences' is the new-word rate over those alone. The i+1 column is measured over them too, since a one-word vocabulary card introduces one new word by construction.
The new-word rate falls from 37.8% to 16.8% by part 5, then holds between 15.7% and 17.4% (median 16.6%) for the rest of the deck.
The notes carry level tags (Level001-Level169): 169 levels, Level001 to Level169, about 25 notes each; 100% of the deck is tagged. The default order does not follow them: in 10% of the pairs of notes from different levels, the higher level comes first (0% would follow the tags, about 50% ignores them). How far into the deck the median note of each level sits: Level001 6%, Level025 20%, Level049 34%, Level073 48%, Level097 63%, Level121 77%, Level145 91%, Level169 5%.
Format: 4,149 of 4,208 measured notes (99%) are sentences of 4 words or more; 59 (1%) are single words or short phrases. Across the whole deck, 57% of the sentences add at most one new word. 4,149 of the sentences are examples on a card that names the word it teaches; there the headword is the card's one new item, so an example passes only if it brings in no other unknown word.
part | notes | new-word rate | sentences | in sentences | <=1 new word | mean words | vocab so far -----+--------+---------------+-----------+--------------+--------------+------------+------------- 1 | 421 | 32.1% | 406 | 32.0% | 34.2% | 6.8 | 896 2 | 421 | 24.2% | 415 | 24.0% | 52.5% | 6.3 | 1,538 3 | 420 | 21.1% | 414 | 21.1% | 52.4% | 7.0 | 2,163 4 | 421 | 20.3% | 416 | 20.2% | 52.6% | 7.0 | 2,771 5 | 421 | 17.2% | 419 | 17.2% | 55.4% | 8.0 | 3,356 6 | 421 | 17.9% | 417 | 17.8% | 61.2% | 7.0 | 3,893 7 | 421 | 16.5% | 415 | 16.5% | 68.0% | 6.8 | 4,378 8 | 420 | 18.1% | 417 | 18.1% | 62.6% | 7.0 | 4,920 9 | 421 | 16.3% | 414 | 16.2% | 67.6% | 6.9 | 5,408 10 | 421 | 16.5% | 416 | 16.4% | 65.4% | 7.5 | 5,941 'sentences' counts the notes in that part holding at least 4 words; 'in sentences' is the new-word rate over those alone. The i+1 column is measured over them too, since a one-word vocabulary card introduces one new word by construction.
The new-word rate falls from 32.1% to 17.2% by part 5, then holds between 16.3% and 18.1% (median 16.9%) for the rest of the deck.
One general-purpose list per language, the same for every deck, so the vocabulary criterion rests on ranks rather than on the model's sense of which words are common.
Frequency list: wordfreq 3.1.1, cleaned (patch de rev261022a), German ('de') list. Entries are compared as dictionary forms; an entry counts as taught when the deck uses it as a spelling or as a dictionary form.
Left out of the list before matching: 1 web addresses and Roman numerals.
Coverage of the most common words on the list:
top 1,000: the deck teaches 973 of 1,000 (97%)
top 2,000: the deck teaches 1,883 of 2,000 (94%)
When the most common words arrive (position as a share of the deck):
ranks 1-100: 100 of 100 taught (100%), median first appearance at 1% of the deck, 100% of them in the first quarter
ranks 101-500: 394 of 400 taught (98%), median first appearance at 10% of the deck, 94% of them in the first quarter
ranks 501-1,000: 479 of 500 taught (96%), median first appearance at 21% of the deck, 61% of them in the first quarter
ranks 1,001-2,000: 910 of 1,000 taught (91%), median first appearance at 32% of the deck, 35% of them in the first quarter
New words in each tenth of the deck, by rank on the list:
part | new words | top 1,000 | top 5,000 | past 5,000 or unlisted
-----+-----------+-----------+-----------+-----------------------
1 | 1,132 | 37% | 71% | 29%
2 | 619 | 39% | 74% | 26%
3 | 566 | 28% | 73% | 27%
4 | 570 | 16% | 67% | 33%
5 | 529 | 9% | 68% | 32%
6 | 557 | 3% | 59% | 41%
7 | 466 | 3% | 62% | 38%
8 | 517 | 2% | 51% | 49%
9 | 470 | 2% | 50% | 50%
10 | 510 | 1% | 36% | 64%
287 list entries count as taught because a field on the card states the word's dictionary form (a 'Verb Infinitive' field, for example), though the studied text shows another spelling of it.
The 40 most common words on the list that the material never uses:
oh (335), ok (399), scheißen (430), sex (472), gebracht (492), drauf (499), sowas (602), dran (656), googeln (670), co (685), entwickelt (688), live (698), läuten (739), drin (761), mm (791), erstmal (793), website (811), arsch (818), benutzt (835), london (853), erwähnt (855), verbunden (862), fürs (864), erinnert (878), ex (895), cool (934), türkei (947), album (1,024), kg (1,083), werten (1,087), final (1,096), arten (1,114), geil (1,157), rum (1,159), van (1,162), vergangen (1,196), sogenannten (1,199), verdammt (1,200), twitter (1,210), song (1,213)
Words that are rare for where they appear: 201 in all, the rarest few per tenth shown (ranked past 5,000 and past 10x the number of words taught before them). Check each against the cards before calling it a vocabulary choice: grammar terminology and pronunciation respellings in explanation fields land here too.
jackett — rank 23,279, at 0% of the deck, after 102 words
berufseinstieg — rank 22,362, at 5% of the deck, after 748 words
kontinental — rank 22,130, at 3% of the deck, after 480 words
unterredung — rank 21,376, at 5% of the deck, after 753 words
wirbelsturm — rank 23,247, at 14% of the deck, after 1,376 words
liebesbrief — rank 21,557, at 12% of the deck, after 1,280 words
multikulturell — rank 21,304, at 13% of the deck, after 1,321 words
traumberuf — rank 21,069, at 13% of the deck, after 1,308 words
föderal — rank 23,064, at 27% of the deck, after 2,158 words
arbeitswoche — rank 22,676, at 28% of the deck, after 2,209 words
erntedankfest — rank 22,400, at 26% of the deck, after 2,072 words
hochschulabschluss — rank 20,657, at 25% of the deck, after 2,036 words
6 distinct words were left out as English explanation text: each is at least 30 times more common in English than in the target language.
712 of the 5,936 distinct words the deck teaches are not among the list's 23,577 entries. That count mixes rare vocabulary with compounds, rare inflections, spelling variants and notation, so none of them are named above as rare; names are already left out.
anki-deck-analyzer "German_Top_4000_vocabulary__1655357470.apkg"
2325a7f33008c9bc…: a code computed from the file's exact bytes. If the same calculation on your copy gives the same code (Get-FileHash on Windows, shasum -a 256 on macOS), you have exactly the deck that was scored.