The Council of Europe published the original framework in 2001 under the name Common European Framework of Reference for Languages, abbreviated CEF. A year or two in, almost everyone began appending an extra R for Reference, and the acronym settled into CEFR. Both labels point to the same document. If you see CEF in older papers, IELTS conversion charts, or a Council of Europe PDF from the early 2000s, treat it as identical to today's CEFR.
The framework was never meant to be a test. It is a descriptive scale. It tells you what a learner can do with a language, broken down into six broad bands and four skills. That's it. No exam, no fee, no certificate of its own. Test providers, schools, ministries, and employers map their own assessments onto the scale, and that mapping is what gives the framework its real-world power.
You'll bump into CEFR levels on job ads, university application portals, EU residence forms, Erasmus paperwork, and the back of language certificates. A line that says "B2 minimum" carries the same meaning in Lisbon, Helsinki, and Bratislava. That cross-border fluency in language requirements is exactly what the Council of Europe was after when it commissioned the project in the 1990s. Whether the document is called CEF or CEFR, you're looking at the same six-step ladder.
This guide walks you through every level, every skill, the Can-Do philosophy, exam crosswalks, immigration use, and the criticisms that keep showing up in academic journals. Read it once and you'll understand why "my CEFR is B1" tells a recruiter far more than "I speak intermediate French."
The ladder splits into three broad bands. A is for basic users, B for independent users, and C for proficient users. Each band has two rungs β A1, A2, B1, B2, C1, C2 β and the gap between them widens as you climb. Going from A1 to A2 takes maybe 100 guided hours. Going from B2 to C1 can take 400 to 600. The scale isn't linear, and pretending it is sets students up for frustration.
At A1 (Breakthrough) you can introduce yourself, order coffee, and ask where the bathroom is, provided the other person speaks slowly and clearly. A2 (Waystage) bumps you up to short transactional exchanges β shopping, basic directions, family talk. B1 (Threshold) is the first level where you can actually function in the language: you handle most situations while travelling, you can describe experiences and dreams, and you can produce a connected stretch of text on familiar topics. Plenty of tourist visas and lower-tier work permits sit right here.
B2 (Vantage) is the workplace ceiling for many migrants. You understand the main ideas of complex text on both concrete and abstract subjects, including technical discussions in your field. You interact with fluency and spontaneity that makes regular conversation with native speakers a real possibility. C1 (Effective Operational Proficiency) means you handle academic and professional language with ease β write structured essays, follow lectures, argue a position. C2 (Mastery) isn't "native speaker" β the document is careful to say so β but it's the level where you read between the lines, catch idiom, and shift register effortlessly.
For a closer look at any single rung, the CEFR level reference page breaks down descriptors in plain language.
A1 β Breakthrough: survival phrases, very slow speech needed.
A2 β Waystage: simple transactions and personal information.
B1 β Threshold: can handle travel and familiar topics independently.
B2 β Vantage: fluent enough for work; understands abstract topics.
C1 β Advanced: full academic and professional use.
C2 β Mastery: precise, idiomatic, near-effortless command.
What makes the framework genuinely useful isn't the six labels β it's the thousands of Can-Do statements sitting underneath them. Instead of vague claims like "intermediate grammar," each descriptor begins with the words "Canβ¦" and describes a real-world task. "Can write very short, basic descriptions of events, past activities, and personal experiences." That's an A2 writing descriptor, and you can imagine exactly what student work would meet it.
This positive, performance-based phrasing was a deliberate break from older deficit models that defined learners by what they couldn't do. It also makes the scale teachable. A B1 writing curriculum maps directly onto B1 Can-Do statements β postcards, emails to friends, short connected paragraphs about familiar topics. Teachers stop asking "have they covered the past perfect?" and start asking "can they recount a weekend trip in a paragraph that reads naturally?"
The 2018 Companion Volume extended the descriptor set significantly. It added mediation (translating, summarising, facilitating communication between speakers), online interaction (forum posts, instant messaging, video calls), and plurilingual or pluricultural competence (drawing on more than one language in the same exchange). These additions push the framework past its old four-skill model into something closer to how people actually use languages in 2026.
Understanding spoken language across registers β announcements, conversations, lectures, broadcasts. Tested with audio at varying speeds and accents.
Decoding written text β signs, articles, correspondence, literature, technical documentation. Speed and inference grow level by level.
Splits into spoken production (monologue) and spoken interaction (conversation). Pronunciation, fluency, range, and accuracy are all tracked.
Producing connected text appropriate to purpose and audience β postcards at A2 through to argued essays and reports at C1-C2.
The framework includes a self-assessment grid you can use without spending a euro. The grid lists each of the four skills (plus spoken interaction and spoken production separately) against all six levels. You read the descriptors row by row and tick the highest one where you can honestly answer "yes, I do that consistently and without effort."
The honesty matters. People routinely overshoot by a full band because they conflate "I once managed this" with "I do this reliably." A useful rule: if you can do the task on a tired Tuesday afternoon in a noisy cafΓ©, you're at that level. If you can only pull it off when you're alert, prepared, and the topic is familiar, you're probably one rung lower.
Free online level tests are everywhere, and most of them are quick multiple-choice quizzes that test grammar and vocabulary. They give you a rough estimate, nothing more. A proper CEFR test includes productive skills β speaking and writing β assessed by a trained rater. Practice quizzes still have value as a first sanity check, especially if you take several from different providers and look at where the scores cluster.
If you want a more rigorous picture, sit one of the placement tests run by Cambridge English, Goethe-Institut, the Cervantes Institute, or the Alliance FranΓ§aise. They're built around CEFR descriptors and they take the productive skills seriously. Our CEFR English test page lists the leading options.
The basic-user band covers the first 200 to 300 guided hours of study for most European languages. A1 work centres on phonetics, the most frequent thousand words, and present-tense survival phrases. A2 introduces past tense, comparison, and slightly more abstract topics like work or holidays. Learners at A2 can complete a simple form, write a short postcard, and follow a slow speaker on familiar subjects, but they still translate mentally and pause often.
This is where most adult learners settle, especially those studying for work or migration. B1 brings genuine independence β you can navigate a country, handle predictable conversations, and produce a connected paragraph. B2 is the workplace threshold cited by most employers and universities. You read journalism, discuss abstract topics, and write structured reports. The B1-to-B2 jump is the steepest of all six and routinely takes 300 to 500 hours of focused study.
The proficient-user band is academic and professional territory. C1 candidates write argued essays, follow university lectures, and adjust register without thinking about it. C2 is described as full operational mastery β the document explicitly avoids calling it "native-like," because native speakers vary wildly. C2 learners read 19th-century novels, catch puns, and shift registers from boardroom to bar without friction. Very few exam candidates ever sit C2 β it's a credential mostly used by translators, linguists, and language teachers.
Most major language certificates publish a CEFR alignment table. Cambridge English built its suite around the framework from the start: KET maps to A2, PET to B1, FCE to B2, CAE to C1, and CPE to C2. IELTS doesn't sit at a single level β its 9-band scale stretches across A1 to C2.
A score of 4.0 lands around A2-B1, 5.5 sits at B1, 6.5 at B2, 7.5 at C1, and anything above 8.5 falls within C2. TOEFL iBT also publishes a chart: 42 hits B1, 72 reaches B2, 95 maps to C1, and a score in the 110s starts approaching C2.
For Spanish, the DELE certificates run from A1 through C2 with one exam per level, all aligned to the framework by the Cervantes Institute. German learners sit TestDaF (which clusters around B2-C1) or the Goethe-Zertifikat series, which mirrors the CEFR levels one for one. The French DELF/DALF exams cover A1 to C2. Italian has CILS, Portuguese has CAPLE, Russian has TORFL. Almost every European language has a state-backed certificate scheme built on CEFR.
Outside Europe, alignment is patchier. China's HSK nominally maps to CEFR but the levels don't line up cleanly β recent HSK 3.0 reforms changed the correspondence yet again. Japan's JLPT sits beside CEFR rather than inside it, though the JLPT N3 is widely treated as roughly B1. The point is to check the chart before assuming your certificate proves a CEFR level. A few exam boards inflate their alignment for marketing reasons.
Most EU member states write language requirements into residence, naturalisation, and work-permit rules using CEFR labels. Germany asks for B1 for permanent residence and naturalisation. France requires A2 for a 10-year residence card and B1 for citizenship. The Netherlands sets A2 for the integration exam, with B1 increasingly common for professional registration. Austria, Belgium, Italy, and Spain all use the framework in similar ways. The list rarely cites exam names β it cites CEFR levels, then publishes a short list of approved certificates that prove them.
Outside the EU, Canada uses the Canadian Language Benchmarks (CLB), a parallel 12-level scale. The two scales aren't a perfect overlay, but published crosswalks are routinely used: CLB 4 sits near A2, CLB 5-6 near B1, CLB 7-8 near B2, CLB 9-10 near C1, and CLB 11-12 near C2. If you're moving from Europe to Canada and your French or English certificate cites CEFR, expect to convert the score for IRCC paperwork rather than swap one for the other.
The UK, post-Brexit, kept CEFR in its visa rules. The points-based system specifies B1 for skilled-worker routes and B2 for some healthcare visas. Australia and New Zealand, by contrast, prefer to cite IELTS bands directly, though the bands themselves carry implicit CEFR equivalents.
Alongside the framework itself, the Council of Europe published a personal document called the European Language Portfolio (ELP). It has three parts. The Language Passport records your levels in every language you speak, with date and certifying body. The Language Biography tracks your learning history β courses, immersion experiences, intercultural moments. The Dossier stores work samples that demonstrate your level β essays, audio clips, project files.
Schools across Europe adopted the ELP enthusiastically in the 2000s, less so by the 2020s. Digital alternatives like Europass have largely absorbed its functions. The Europass CV template now includes a CEFR self-assessment table for each declared language, which is almost certainly where you'll meet the framework if you're job-hunting in Europe. Filling that table in honestly β and in CEFR terms β signals to recruiters that you understand the scale and aren't bluffing.
One genuinely useful idea the portfolio captured was plurilingualism. The document treats every learner as someone who already brings linguistic resources to a new language. A Polish speaker learning Italian draws on Slavic cognates, school English, maybe German from a holiday job. The portfolio asks you to acknowledge that mix instead of pretending each language sits in a sealed box.
The framework isn't unchallenged. Applied linguists have raised several recurring complaints. The first is scale granularity. Six levels stretching from absolute beginner to near-bilingual is too coarse for genuinely diagnostic feedback. A learner at the top of B1 and one at the bottom of B2 might sit a centimetre apart on the ladder, yet they need very different teaching. The 2018 Companion Volume responded by introducing Pre-A1 and a handful of "plus" levels (A2+, B1+, B2+), but the core six-step structure stayed.
The second criticism is cultural specificity. The descriptors lean on text types and communication norms that are heavily European and heavily academic β letters of complaint, abstract argumentative essays, classroom-style discussions. Researchers working with learners from non-Western traditions argue the scale doesn't always fit the language ecology those learners live in. Some Asian and African scholars have proposed parallel local frameworks rather than retrofitting CEFR to contexts it wasn't built for.
The third concern is misuse by gatekeepers. Immigration ministries, employers, and admissions offices treat CEFR labels as hard cut-offs when the document itself describes them as overlapping bands. A candidate scoring 6.4 on IELTS β clearly in the B2 band β gets refused for a B2-only opening because they need 6.5. The framework was never designed to support that level of precision. The Council of Europe's own guidance asks users to treat the levels as broad descriptors, not certification thresholds. In practice that warning is widely ignored.
A final, milder complaint is that the framework remains under-documented for less commonly taught languages. The full descriptor set works best for English, French, German, Spanish, and Italian. Speakers of Welsh, Albanian, Maltese, or Faroese have far fewer aligned materials, fewer certifying exams, and fewer trained raters. The framework is universal in principle and patchy in practice.
There's also a quieter debate about how the framework is taught versus how it is tested. A class can hit every B1 descriptor while a single high-stakes exam still produces a B2 pass for a student whose speaking is shaky. Conversely, classroom-shy students sometimes underperform their actual ability on speaking-heavy exams.
Researchers point out that the CEFR was built to describe communicative competence over time, not to be pinned to a 90-minute paper. When ministries and employers ignore that nuance, learners with strong receptive skills but weaker productive ones β and vice versa β get sorted into the wrong band.
Two big shifts are reshaping the framework in the mid-2020s. The first is the slow integration of plurilingual and mediation descriptors into mainstream exams. Cambridge has piloted mediation tasks; the new Goethe-Zertifikat C1 includes a summary-and-paraphrase task that explicitly maps to the 2018 Companion Volume. Expect more exams to follow within five years.
The second is the slow recognition of online interaction as a distinct competence. The framework now has descriptors for forum discussion, instant messaging, video-call etiquette, and collaborative document editing. These show up first in workplace assessments β corporate language tests increasingly include screen-share scenarios β and they will land in mainstream exams next. The CEFR you learned in 2015 already looks different from the one being used today.
If you take one practical lesson from the framework, make it this: stop thinking about language in terms of vocabulary lists and grammar coverage. Think about what you can do, what you want to do, and what level the people you'll talk to need from you. That's the question CEFR was built to answer. For a quick self-test, the CEFR levels explained overview pairs each band with example tasks you can try right now.
Yes. CEF was the original abbreviation used in the 2001 document; CEFR is the longer form that became standard a couple of years later. The two acronyms refer to the same Council of Europe framework, so any chart or rule citing CEF maps one-for-one onto CEFR.
It varies, but Cambridge and the Goethe-Institut publish broadly similar guidance. Roughly 100 to 200 guided hours per level at the lower bands, rising to 400 to 600 hours to move from B2 to C1. Going beyond C1 to C2 is often described as open-ended.
Yes. CEFR exams are independent β you can sit any level without proving you've passed the one below. Most candidates take a placement test first to confirm they're ready, since a fail at B2 costs the same as a fail at B1.
No. The Council of Europe is explicit: C2 is full operational mastery, but native speakers vary enormously in vocabulary, register, and literacy. C2 is the ceiling of the scale, not a claim about being indistinguishable from a native.
The official IELTS-CEFR alignment maps 4.0-5.0 to B1, 5.5-6.5 to B2, 7.0-8.0 to C1, and 8.5-9.0 to C2. Most universities and immigration bodies cite the band, not the CEFR level, in their cut-offs.
Yes β widely. Cambridge English exams used worldwide are CEFR-aligned, Canada cross-references CEFR against CLB for immigration, and many universities in Asia, Latin America, and the Middle East cite CEFR levels in admissions. Coverage and alignment quality vary.
The Council of Europe self-assessment grid is free and printable. For something more interactive, Cambridge, Goethe-Institut, Cervantes, and Alliance FranΓ§aise all offer free online placement tests aligned to CEFR. Our CEFR test page lists the most reliable free options.
Originally yes β the 2001 document built around the four classical skills. The 2018 Companion Volume added mediation, online interaction, and plurilingual or pluricultural competence as further descriptor sets, so a fully up-to-date CEFR test now reaches beyond the original four skills.