What CEFR levels actually mean
Not a score, a description
The most common misunderstanding about CEFR levels is that they are a grade. B1 sounds like a middling mark; C1 sounds like an A. That reading makes the levels useless, because it turns a description of behaviour into a rank.
CEFR bands are can-do descriptors. Each one is a statement about what a person can actually do in a language, in situations, under normal conditions. “B1” does not mean you know a certain number of words. It means you can deal with most situations that come up while travelling, working or studying in the language, produce connected speech on familiar topics, and describe experiences, opinions and plans while giving reasons for them.
That difference is practical, not academic. A score tells you where you sit. A descriptor tells you what to practice, because it names the thing you cannot yet do.
The bands, in terms of conversation
Here is the scale as it applies to speaking, which is the skill this blog is about.
A1 is basic, concrete, everyday exchange. Introducing yourself, asking and answering simple personal questions, handling survival situations — ordering food, asking a price, asking where something is. Speech from the other side needs to be slow and clear, and support in your own language is expected and normal. The grammar underneath is the present simple, to be, articles, possessives, question words, can, and going to for near-future plans.
A2 is routine tasks and short social exchanges. Narrating past events, describing what you usually do against what you are doing now, handling shopping, travel and appointments, and telling a short personal story with basic connectors. The distinguishing feature is connected sentences: not single utterances, but two or three joined with because, but, so, then.
B1 is the first band where you are genuinely independent. You can deal with most situations that arise while travelling or at work, justify an opinion, talk about hypotheticals, handle a workplace exchange, manage a complaint politely, and — importantly — recover when a conversation breaks down, by asking for clarification instead of stopping. Grammar-wise this is present perfect against past simple, conditionals, relative clauses, reported speech, and the passive.
B2 is fluency and spontaneity with limited strain on either side. You can develop and defend an argument with supporting detail, negotiate, summarize a source, chair a meeting, shift register between formal and casual, and paraphrase your way around a word you cannot retrieve rather than freezing. This is usually the band people mean when they say they want to be “fluent.”
C1 is precision. Not more range so much as better control: arguing a complex position with nuance, handling a diplomatic or difficult conversation, using idiom accurately rather than approximately, conveying hedging and irony, and sustaining a long turn with proper signposting. The grammar work at this level is refinement rather than acquisition — residual article and preposition errors, collocation accuracy, aspect precision in extended narration.
Above that is C2, which is a different project again and further than a working adult usually needs to go.
Why the descriptors are better goals than the labels
“I want to reach B2” is a weak goal, because it does not tell you what to do on Tuesday. “I want to be able to give a status update in a meeting, raise a blocker, and propose next steps without rehearsing it first” is a strong one, and it happens to be a B2 objective.
This is why a course built on CEFR should be a list of can-do objectives, not a band label with content hung off it. A band is a bundle of specific abilities: narrate the past, justify an opinion, negotiate, repair a breakdown. Each is separately practisable, and each is separately assessable — you either did it in a conversation or you did not.
It also explains the common experience of feeling stuck between levels. You are not stuck; you have a band where most objectives are solid and a few are not, and the few are the ones nobody has been targeting because “improve my B1” is not an instruction anyone can act on.
What “ready for the next level” should mean
A serious answer to “am I B2 yet” needs a rule stated in advance, otherwise it is a vibe.
The rule worth using is about the objectives, not about a test score. Each objective in a band sits in one of three states: being introduced, being practiced, or mastered. Moving up a band means the core objectives of the current one are mastered and the remainder are at least in active practice. For A1 that means the everyday-exchange objectives are solid and the rest are underway; for B1 it means the experience, opinion, hypothetical, workplace, planning, reported-speech, description and complaint objectives are done.
Alongside that sit qualitative exit criteria that describe the same thing in prose. Leaving A1 means sustaining a simple five-minute conversation about yourself, your family and your routine, with frequent but non-blocking errors, and needing only occasional support in your own language. Leaving B2 means interacting fluently and spontaneously, arguing and negotiating effectively, shifting register appropriately, and self-repairing without help, with errors that are occasional and do not impede communication.
Notice what is in both formulations: errors are allowed. “Frequent but non-blocking” is an A1 exit criterion, not a failure state. A learner who waits until they stop making mistakes before considering themselves A2 will wait forever, and will spend the wait not speaking.
How long a band takes
Curriculum arcs at the lower bands are shorter than at the higher ones — roughly ten weeks of structured work for A1 and A2, and closer to twelve for B1, B2 and C1. That asymmetry is not arbitrary. Early bands have less material and more obvious wins; upper bands ask for control and nuance across a much wider surface, and control takes repetition rather than exposure.
Those are arcs of a curriculum, not predictions about a person. Actual pace depends on how much you speak per week, which is the variable that matters most and the one entirely under your control.
Where a level estimate comes from, and how much to trust it
A level from a single conversation is a starting point, not a verdict. It is worth stating that explicitly, because a label handed to a learner in the first session tends to stick emotionally regardless of its accuracy.
A defensible estimate carries three things: the band, a confidence rating, and the evidence behind it — specific things you said or failed to say. Confidence matters because a first estimate made on a short, nervous conversation genuinely is less reliable than one made after several sessions, and a system that reports the band without the confidence is hiding the useful half.
The estimate then gets refined over the first few lessons as more evidence arrives. This is also why a level should be allowed to move down as well as up. A first conversation where you happened to be on comfortable ground can read a band high, and every subsequent lesson pitched at that band will feel like drowning for reasons that are not your fault.
What this changes about how you practice
Three things follow from taking the descriptors seriously.
Pick the objective, not the band. “Practice justifying opinions” is a session. “Practice B1” is not.
Practice the repair skills, which are objectives in their own right. Asking for clarification, paraphrasing around a missing word, and recovering from a breakdown are explicitly part of B1 and B2 — they are not what you do when the lesson goes wrong, they are part of the lesson.
And accept that the band moves on evidence of production, not on study time. Reading about the present perfect does not move a speaking objective. Producing it, unprompted, in a conversation where it was the right choice, does.
The method page describes how a CEFR skeleton sits under a personalized plan, and the closed learning loop post covers how objective states get updated from one lesson to the next.