How long does it take to learn German?

There are two sources worth citing on this, and almost every page that quotes them gets at least one thing wrong. Here is what they actually say.

The Goethe-Institut figures

The Goethe-Institut publishes guidance in Unterrichtseinheiten (UE), where one UE is 45 minutes of instruction.

These are cumulative totals to reach the level, not the cost of each step. This is the most common misreading, and it matters: adding the rows together to get "A1 through B2" roughly doubles the real figure. Getting from zero to B1 is somewhere in the 350-650 UE band in total - not 80-200 plus 200-350 plus 350-650.

Note also how wide the bands are. B1 spans 350 to 650 - nearly double from one end to the other. That range is doing real work, and Goethe says why: individual factors, prior language-learning experience and your first language all move the number.

The FSI figure, and the one that is wrong

The US Foreign Service Institute groups languages by how long its students take to reach professional working proficiency. German is Category II: approximately 36 weeks, or 900 class hours.

If you have seen "750 hours" for German, that figure is wrong. 750 hours is the top of FSI's Category I range (24-30 weeks, 600-750 hours), which covers languages like Spanish, French and Dutch. German sits one category above them - it gets its own category largely because of the case system and word order. The 750 number circulates widely and is simply a misattribution.

For context, Category III is around 1100 hours and Category IV - Arabic, Chinese, Japanese, Korean - is around 2200.

FSI is explicit that these are averages from observation, and that actual time varies with "natural ability, prior linguistic experience, and time spent in the classroom."

Why you should not put these two numbers side by side

This is the part almost everyone gets wrong, and it is worth understanding rather than just accepting.

They measure different targets. Goethe measures against the CEFR (A1-C2). FSI measures against the ILR scale, aiming at Speaking-3/Reading-3, "General Professional Proficiency". Neither body publishes a mapping between CEFR and ILR. Any equivalence you see between them is someone's inference, not an official one.

They use different units. FSI's 900 are 60-minute hours. Goethe's UE are 45 minutes. So FSI's 900 hours is about 1,200 UE - which is above Goethe's C1 band, not equal to it, even though "900 hours" and "800-1000 UE" look similar at a glance.

The honest summary is that these are two different institutions answering two different questions, and the resemblance between their numbers is a coincidence of arithmetic.

What neither of them tells you

Neither publishes a conversion between self-study and classroom hours. FSI's figures are explicitly class hours with intensive instruction and small groups. Goethe's are explicitly Unterricht.

So when an app tells you that fifteen minutes a day gets you to B1 in some number of months, that claim is not derived from either of these sources, because neither source supports it. We are not going to offer a conversion ratio either, for the same reason - we do not have one that is honest.

What can be said without inventing anything: the FSI numbers assume conditions almost no independent learner has, and self-study is generally less efficient per hour than intensive classroom instruction, not more.

What actually moves the number

Since the ranges are so wide, the interesting question is what puts you at the low end.

• Languages you already have. Dutch or a Scandinavian language helps enormously. English helps more than people expect, because a large amount of German vocabulary is transparent once you notice the sound correspondences.
• Intensity beats duration. Two hours a day for three months beats twenty minutes a day for eighteen. Spacing helps retention, but very thin daily contact means you spend most of your time re-learning what decayed since the last session.
• Production versus recognition. Time spent recognising German is much less expensive than time spent producing it - and it is production that exams and conversations test. Hours logged reading comprehension are not fungible with hours spent forming sentences.
• Whether anything brings grammar back. A course that explains the Dativ once in chapter 7 has not taught you the Dativ; it has shown it to you. The re-learning cost of everything you met once and forgot is invisible in these hour estimates and is often where the time actually goes.

That last point is the one worth optimising, and it is the reason spaced repetition exists. If you want to see the scope of what B1 or B2 actually contains, the grammar library is readable without an account, and the vocabulary is ordered by frequency so you can see what the first thousand words look like.