The Common European Framework of Reference for Languages, or CEFR, is a six-level scale describing what a learner can do in a foreign language, from A1 at the beginner end to C2 at the top. It was published by the Council of Europe in 2001 after two decades of development, and it is now the default vocabulary for language ability across Europe and much of the world. Its defining feature is that each level is set by can-do statements rather than by grammar covered or tests passed.
What the CEFR actually is#
The framework grew out of Council of Europe work stretching back to the 1970s, including van Ek's Threshold Level (1975), which specified the minimum language needed to function independently in a foreign country. That specification became B1. The full framework was published in 2001, and a Companion Volume with New Descriptors followed in 2018, updated in 2020, adding scales for mediation, plurilingual competence, sign language, and a Pre-A1 level.
It is a reference framework, not a curriculum. It does not tell a teacher what to teach or in what order, and it was designed to work for any language. That choice explains most of what people find frustrating about it — no word lists, no grammar sequence, no hours — because anything language-specific would have broken the neutrality.
Can-do statements, not test scores#
Every CEFR descriptor is phrased as something a learner can do. The B1 global descriptor says a learner "can understand the main points of clear standard input on familiar matters regularly encountered in work, school, leisure," and "can produce simple connected text on topics which are familiar or of personal interest."
Three properties follow. The descriptors are criterion-referenced: you are measured against a description of performance, not against other candidates. They are positively worded: what you can do, not what you fail at. And they are skill-specific, with separate scales for listening, reading, spoken interaction, spoken production, and writing — which is why the self-assessment grid has five columns rather than one number, and why "I am B1" compresses something more complicated.
The three bands#
The letters matter more than most learners realize. A, B, and C mark shifts in kind, not just degree.
A — Basic User. You handle predictable, concrete situations with rehearsed language. Communication depends heavily on your interlocutor slowing down and helping.
B — Independent User. You function without that help. B is where the language stops being a set of memorized situations and becomes a tool you can point at new problems.
C — Proficient User. You handle abstract, implicit, and stylistically demanding material, and can operate professionally or academically in the language.
The CEFR also recognizes intermediate criterion levels — A2+, B1+, B2+ — for the long stretches inside a band where progress is real but has not crossed the next threshold. The B1-to-B2 span is wide enough that it is where most learners stall, as covered in the intermediate plateau.
The six levels in full#
| Level | Band | Core claim | Typical reading | Approximate vocabulary (estimates vary) |
|---|---|---|---|---|
| A1 | Basic | Familiar everyday expressions and very basic phrases; introduces self, asks simple personal questions | Signs, menus, labels, single sentences with support | ~500–1,000 words |
| A2 | Basic | Sentences on immediate topics — personal details, shopping, employment; routine exchanges | Short simple texts, personal letters, predictable notices | ~1,000–2,000 |
| B1 | Independent | Main points of clear standard input on familiar matters; copes while traveling | Straightforward factual texts, graded readers, simplified news | ~2,000–3,000 |
| B2 | Independent | Complex text on concrete and abstract topics; interacts without strain on either side | Newspaper articles, contemporary prose, professional material in a known field | ~3,000–5,000 |
| C1 | Proficient | Demanding longer texts, including implicit meaning; expresses ideas without obvious searching | Literary fiction, dense journalism, academic writing | ~5,000–8,000 |
| C2 | Proficient | Virtually everything read or heard; reconstructs arguments; finer shades of meaning | Anything, including wordplay, dialect, and technical material outside your field | ~8,000–16,000 |
The vocabulary column is the least reliable part of that table.
What reading looks like at each level#
Reading is the skill where the levels are easiest to feel, because you can hold up a text and check. Paul Nation's coverage research is the useful lens: comfortable independent reading needs roughly 98% of the running words on a page to be known, with about 95% workable when support such as a glossary is at hand.
At A1 and A2, native text sits far below that threshold, which is why graded readers exist. At B1 simplified news becomes readable and native text becomes possible with heavy support. At B2 a newspaper article in a familiar domain crosses into comfortable territory, though fiction with idiom and voice often does not. At C1 the remaining gaps are stylistic and cultural rather than lexical. Below B2, then, you should either read graded material or read native material with part of the vocabulary handled for you.
That second option is the diglot weave method — reading text you already understand with a controlled share of target-language words woven in, so coverage stays above the comprehension threshold while difficulty climbs. It is the approach LingoBlend is built on, and it exists because the gap between A2 vocabulary and native text is too wide to jump directly.
Why two people at the same level look different#
This is the most common source of confusion, and it has three causes.
Uneven skill profiles. Someone who learned by reading may read at B2 and speak at A2; someone who learned by living abroad is often the reverse. The headline level averages away a difference that dominates how the person seems in conversation.
Domain dependence. The descriptors repeatedly qualify with phrases like "in his or her field of specialization." An engineer at B2 may read technical papers comfortably and be lost in a restaurant review.
Test date drift. A certificate records performance on one day, most carry no expiry, and ability decays without use — so a B2 from six unused years ago is a historical fact rather than a measurement.
Levels, exams, and hours#
Major exam suites are aligned to the framework: Cambridge English (A2 Key through C2 Proficiency), the Goethe-Zertifikat series, DELE for Spanish, and DELF/DALF for French all map onto CEFR levels directly. Others do not. The Japanese JLPT uses its own N5-to-N1 scale and is not formally CEFR-aligned; mappings you see online are informal approximations.
Exam boards publish rough guided-learning-hour estimates, most commonly for English, ranging from a couple of hundred hours to reach A2 to something over a thousand cumulative hours to approach C2. Treat those as order-of-magnitude figures. They vary enormously with the distance between your native language and the target, with study intensity, and with how much input you get outside class. Language-specific timelines are covered in posts like how long it takes to learn Spanish.
Vocabulary estimates and why the CEFR gives none#
The framework specifies no word counts. The numbers in the table above come from language-specific research projects and exam-board profiling work, and sources disagree substantially — partly because they count different things. A "word" might mean a word family, a lemma, or a distinct written form, and those counts can differ by a factor of two on the same text. Treat any single figure quoted for a CEFR level as one estimate among several, not a target. The counting problem is worked through in how many words it takes to be fluent in Spanish.
Using the framework instead of arguing with it#
The levels are most useful as a way to pick material. Find the level where you understand nearly everything, read a great deal at it, and step up when it stops being challenging. If you are approaching the B band, what B2 level means sets out the threshold you are aiming at, and LingoBlend's languages page lists where each one starts.