What CEFR levels actually mean when you are reading
A1 to C2 explained in terms of what you can genuinely get through, why the vocabulary numbers you see quoted vary so much, and why your reading level is probably higher than your level.
Every language app, textbook and course now labels itself A2 or B1. Almost none of them tell you what that means for the thing you actually want to do, which is pick up something written for adults and get through it.
The labels come from the Common European Framework of Reference, and they are more useful than their reputation suggests. But they describe a person, and reading happens between a person and a particular text, so the label alone will never tell you whether you can read the thing in your hand.
Here is what each level means in practice, what the vocabulary research says, and why the numbers you see quoted vary so wildly.
The six levels, as a reader experiences them
The official descriptors for reading are reasonably concrete. Paraphrased, and with what they mean in practice:
| Level | What you can read | Vocabulary, roughly |
|---|---|---|
| A1 | Names, single words, very simple sentences. Signs, labels, a menu. Not yet continuous text. | under 1,500 |
| A2 | Short, simple texts. Finding one predictable fact in a timetable or an advert. A short personal message. | 1,500 to 2,500 |
| B1 | Everyday and work language at length. Someone describing events and feelings. This is where continuous reading starts being possible. | 2,750 to 3,250 |
| B2 | Articles and reports where the writer takes a position. Contemporary novels. The level where the real world opens up. | 3,250 to 3,750 |
| C1 | Long, complex factual and literary texts, and you notice style rather than only meaning. Specialist writing outside your own field. | 3,750 to 4,500 |
| C2 | Effectively anything, including abstract and structurally difficult writing. | 4,500 to 5,000+ |
The single most useful line in that table is B1 to B2. That is the stretch where reading stops being an exercise and starts being reading, and it is also where most learners stall, because it is the point at which courses stop and authentic material has not yet become comfortable.
Why the vocabulary numbers should be treated gently
Those figures come from Meara and Milton’s work using the XLex test, and they are the most commonly cited mapping. They are also softer than they look, in three ways worth knowing before you quote them at anyone.
The test has a ceiling. XLex measures out of 5,000, so “C2 is 4,500 to 5,000” partly reflects where the ruler stops. Educated native speakers know something closer to 20,000 words. C2 is not the top of vocabulary knowledge; it is the top of that scale.
The same level means different vocabulary in different places. When Milton and Alexiou tested over 500 learners of English, French and Greek across five populations, B1 learners scored anywhere from about 2,200 to about 3,300 depending on the language and the group. That is a thousand-word spread inside a single label. The standard deviations overlap between adjacent levels, which is a polite way of saying that a strong A2 and a weak B1 are not distinguishable by vocabulary size.
The units differ between studies. These counts are of words in frequency bands. The coverage research uses word families, which group a word with its inflections and derivations, so one family may cover several counted words. Numbers from the two traditions cannot be compared directly, and a great deal of confident internet advice does exactly that.
So treat the column as an ordering rather than a measurement. Levels go up, vocabulary goes up, and the gaps between them are real. The specific integers are not.
Your reading level is probably higher than your level
The CEFR describes several skills, and most people are not the same level in all of them. Recognising a word when you see it is far easier than producing it, and receptive vocabulary is generally a good deal larger than productive vocabulary.
This has a practical consequence that most learners get wrong: if you have been assessed at B1 overall, you can probably read at the top of B1 and reach into B2, particularly on a subject you know. Someone who has been told they are B1 and only ever reads B1 material is reading below themselves.
The reverse is true for listening, where most people are behind their reading. That gap is normal and it is worth closing deliberately rather than being embarrassed by.
The level of a text is not a property of you
This is the important part, and it follows from how coverage works.
Whether you can read something depends on how many words in that text you already know. You do not have one reading level; you have a different one for every text in the world. A B1 learner who follows cycling can read cycling reports that would defeat a B2 learner who does not, because the vocabulary that matters is a small recurring set they already own.
Which means the label is a starting point for choosing material, not a verdict:
- Use the level to pick a shelf, not to pick a book. It tells you roughly where to look.
- Then let the text decide. Read a paragraph. If you are stopping constantly, it is too hard regardless of what either of you is labelled.
- Bias downward when you are starting and upward once you are comfortable. Reading slightly below your level is pleasant and builds volume. Reading well above it builds nothing, because you stop.
If you do not know your level
You do not need a test. Two questions get you close enough to start:
Can you read a short news article on a familiar subject and come away knowing what happened, even with gaps? If yes, you are at least B1 and should be reading real material with support, not graded readers.
Can you get through a page of a contemporary novel without stopping more than a couple of times? If yes, you are into B2 and the constraint on what you read is now interest rather than level.
If neither, start with graded material at A2 or B1 and let it carry you. That is not a lesser activity. It is the same activity, at a coverage where it works.
Sources: Council of Europe (2001), Common European Framework of Reference for Languages, self-assessment grid for reading; Meara, P. and Milton, J. (2003), XLex: the Swansea Vocabulary Levels Test; Milton, J. (2010), The development of vocabulary breadth across the CEFR levels, EUROSLA Monographs 1, including Milton and Alexiou’s cross-language data.