Free CEFR placement, 12-36 items

English level test that stops when it knows

Six blocks of six multiple-choice items, one block per CEFR level, and the run walks up and down the ladder rather than through a fixed list: it opens at B1, drops a level if that block does not hold, climbs while blocks keep falling, and ends at the first one you do not pass with four right out of six. Free, no account, nothing held back. The result leads with the block-by-block record and puts the level second, because the level is this page’s own reading of that record and the record is the part that is measured.

100% free
No signup
6 blocks of 6 items
A1 to C2
Stops at your ceiling

Six blocks of 6 items, one block per CEFR level, and the run walks rather than marches. It starts at B1: pass the block and the next one up is opened, fail it and the walk drops a level until something holds. It stops at the first block you do not pass, which is between 12 and 36 items in total.

The six blocks, and what each one is made of

A1Breakthrough — present of be, simple negation, articles, the first thousand words
A2Waystage — past simple, comparatives, since and for, verbs that travel badly from other languages
B1Threshold — first conditional, reported speech, dependent prepositions, verb plus gerund
B2Vantage — third conditional, future perfect passive, fixed collocations, connector syntax
C1Effective operational proficiency — inversion after no sooner, the mandative subjunctive, pied-piped prepositions, near-synonym choice
C2Mastery — concessive however, lest plus subjunctive, low-frequency idiom, evaluative understatement

Each block holds 6 items: two on form, two on word choice and two on a short piece of text. The mix is fixed at every level so that a block record is comparable up and down the ladder — a level failed on vocabulary and a level failed on grammar are the same 4 of 6 threshold.

Starting at B1 is the shorter route and reaches the same answer, because a block passed on the way up is evidence about the blocks below it. Starting at A1 asks every level in order and takes all 36 items — worth it only if you want to see the whole ladder. Neither route is timed.

How the placement walk works

A basal, then a ceiling, then the level between them — the procedure an examiner uses with a printed battery, run on six CEFR blocks.

Open at B1, or ask for the whole ladder
The default entry point is the B1 block, which is the middle of the framework and the level most people who search for a placement turn out to be near. Choosing A1 instead runs every block from the bottom in order, all 36 items, and reaches the same answer more slowly. Neither route is timed and neither shows you whether an answer was right while the run is going, because a verdict after each item teaches the grammar the next item is about to test.
Answer six items per block, pressing 1 to 4 or tapping
Every block holds the same mix: two items on form, two on word choice and two on a short piece of text — a sign, an email, a returns policy, a review. The block closes on six answers and is passed at four of them. An item you were not looking at, because the tab went behind something, is put back on the pile rather than marked wrong, so a block is always decided on six real answers.
Read the block table before the level
The result opens with which blocks were tested, in the order the walk took them, what you scored in each and what each one was testing. The level sits above it as a single band with the framework's own description of what someone at that level can do. Under the table sit the three sub-scores — form, word choice, text — and then every item you saw with a sentence on why the key is the key, which is the part worth copying out.

Technical specifications

Items and blocks	36 items in the bank, arranged as six blocks of six — one block per CEFR level. Each block is 2 grammar, 2 vocabulary and 2 short-text items, so a block record means the same thing at A1 as at C2
Length of a run	12 items at the shortest, 36 at the longest. Opening at B1 and failing it costs two blocks if A2 holds and three if it does not; opening at B1 and clearing everything above costs four blocks. Asking for the whole ladder from A1 always costs six
Pass mark	4 of 6 in a block. With four options per item the guessing floor is 25%, and the probability of reaching 4 of 6 by guessing alone is 3.8% — which is why the walk moves on a whole block rather than promoting on a single right answer
How the walk moves	Fail the opening block and it drops one level at a time until a block is passed or A1 runs out; pass it and it climbs until a block fails or C2 is cleared. A level already behind the walk is never re-tested, because a level below a pass and above a fail has already been decided
What the level is	A band this page assigns to a block record, not a calibrated score conversion. Council of Europe (2001), Common European Framework of Reference for Languages: Learning, Teaching, Assessment, Cambridge University Press defines the levels; the alignment of "highest block passed" to a level is ours, and the can-do line beside the result is our summary of the framework's global scale rather than a quotation of it
What voids an item	An answer arriving before its item was painted, and any item the tab was hidden or unfocused during. A voided item goes back on its block's pile and is asked again, so the block still closes on six answers. Reading is untimed here, so nothing is voided for arriving quickly
Skills not covered	Speaking, writing, listening and spelling. Nothing here is produced by you except a choice among four options, so the run measures recognition only. A CEFR level from a certificate rests on four skills, two of them rated by a human
What leaves the page	Nothing, unless you press copy — which puts the block table, the level and the run seed on your own clipboard as plain text. The 36 items ship inside the page, so a run makes no request to any server

Frequently asked questions

I passed B1 and failed B2. Why did it not ask me anything else?

Because the two blocks together already answer the question the walk was asking. A pass at B1 puts a floor under you and a fail at B2 puts a ceiling on you, and there is no level in the framework between them for a third block to explore. Adding items at B1 would make the B1 verdict more certain, which is a different question from where you sit; adding items at A2 would test a level the B1 pass has already cleared. Stopping is not the walk giving up — it is the walk having nothing informative left to ask.

Is this the level I would get from a real exam?

No, and the gap is mostly not about difficulty. A certificate is awarded on four skills: IELTS runs about two hours forty-five minutes of listening, reading and writing plus a separate speaking interview of eleven to fourteen minutes; Cambridge C1 Advanced runs about four hours with a paired speaking test; TOEFL iBT is about two hours and includes both spoken and written responses. Two of those skills are production, and production is where most learners sit a level below their recognition. This page asks you to choose among four options thirty-six times at most, in your own room, with nobody checking who is typing.

The items got easy again halfway through. Did I do badly?

You started at B1 and did not pass it, so the walk went down rather than up. That is the descent phase, and it is the half of the procedure that most placement tests skip: a test that only climbs can tell you where you stopped but not where you are secure. Dropping to A2 and passing there is a more useful result than a B1 block scraped at three of six, because it identifies a level you hold rather than one you were near.

Why six items a block instead of ten or twenty?

Six is the smallest block on which a 4-of-6 pass mark still keeps guessing out. Run the arithmetic on four options and a block falls to a pure guesser 3.8% of the time; drop to four items with a 3-of-4 mark and that rises above 5%, and the walk starts promoting people through levels on luck. Going the other way, ten items a block would make each verdict firmer and would triple the length of a run that is trying to find a level rather than certify one. The precision that matters here is between levels, not within them.

Two of the four options look grammatical to me. Which do I pick?

Pick the one that a careful writer would use, because that is what every key here is set to. Several items are built precisely on options that are grammatical in isolation and wrong in the sentence: a connector that needs a noun phrase where a clause follows, a comparative formed the long way round when the short way exists, a present perfect against a finished time expression. Where an item genuinely admits two answers it is a fault in the item rather than a hard item, and the review at the end shows the reasoning for every key so you can see which of the two it is.

Does the level mean anything if I take it twice?

Less than you would like, because the second run draws the same 36 items in a different order. Once you have seen an item and read why its key is the key, answering it again measures memory of this page rather than English. If you want a second reading, leave several weeks between runs or use the sub-scores rather than the level: the split between form, word choice and short-text items is the part that survives a repeat, because knowing one item's key does not tell you the next one's.

Where does the CEFR level itself come from?

From the Council of Europe, which published the framework in 2001 and extended it with a Companion Volume in 2020. The framework describes six levels in terms of what a learner can do — hold a conversation while travelling, follow an argument in their own field, catch what a text implies rather than says — and deliberately does not define them as score bands on any test. Every mapping from a score to a level, including this page's, is a decision the test makes for itself, which is why the block record is printed above the level and not below it.

What a placement estimate is, and what a certificate is

A placement test and a certification exam answer different questions and are built differently as a result. Placement exists to route someone into the right class or the right material with the fewest items possible, which is why it stops the moment it knows and why nobody expects it to be defensible against an appeal. Certification exists to produce a claim a third party will act on, which is why it takes four hours, covers all four skills, is run under supervision and is scored by people who were trained on a rating scale. The measurement is different in kind, not merely in length: this page never asks you to produce a sentence, so it has no evidence at all about the half of your English that other people actually hear.

The awkward part is the level itself. Council of Europe (2001), Common European Framework of Reference for Languages: Learning, Teaching, Assessment, Cambridge University Press sets out six bands as descriptions of what a learner can do, and pointedly not as score ranges — the framework has no test attached to it. Every exam that reports a CEFR level has therefore made its own alignment decision, and published alignments disagree about where one band ends and the next begins. CEFR levels are descriptors of what a learner can do, not score bands. Every mapping from a test score to a level is that test's own calibration decision, and different published mappings disagree about where B2 begins. That is why this page prints the block record first, names the pass mark it used, and describes the band in the framework’s own terms rather than converting anything into a percentile — there is no population behind these 36 items to be a percentile of. The same reasoning applies to the reading level test, which reports grade bands and names the readability formula behind each one instead of asserting a grade.

Where placement tests go wrong most often is the ceiling. A test whose hardest item is B2 cannot distinguish B2 from C2, so it hands a C1 speaker a B2 and a C2 speaker the same B2, and neither of them learns anything. This one has a C2 block, and if you clear it the result says the ladder ran out rather than pretending C2 is a measurement — the six items above C1 here turn on concessive however, lest with an uninflected verb, understatement that reverses in three words, and one fixed phrase that four real words compete for. Clearing them says the ladder had nothing left, which is a fact about the ladder. If the block that broke was a vocabulary block rather than a grammar one, the synonym test and the vocabulary test size that half on their own; if it was the short-text items, the reading comprehension test separates inference from detail. And if what you want is not a level at all but a diagnosis of which grammatical relation keeps catching you out, the sentence completion test scores five kinds of blank separately and reports no level whatsoever, and the spelling test dictates words aloud and sorts your errors by the rule that produced them.

Nothing on this page is scored against a clock, and no figure in the result is a duration. That is a design choice with a cost worth naming: a slow reader and a fluent one who reach the same block record are reported the same way here, even though a placement interview would hear the difference in a sentence. The trade is deliberate — a clock on a reading item measures the reading, not the grammar — but it means the level below describes what you can work out, not how readily it comes.

This is a measurement exercise, not a clinical assessment. It reports what you did on this page against a stated reference and nothing more — it cannot establish dyslexia, a language disorder or a specific learning difficulty. Only a qualified professional, working with more than a browser, can make that judgment.

It is also not a qualification. Nothing issued on this page can be verified by anyone, because nothing is issued: the level is computed in your own browser from six answers per block, and no record of it exists anywhere once you close the tab.

Where the 36 items are marked

Every number on this page is worked out by JavaScript running in the tab you are reading it in. Your answers, your reaction times and your score are never uploaded, logged or kept — which is also why the test carries on working after you disconnect from the network, and why nothing here can be held back behind an email address.

The item bank, the keys and the walk rule are all inside the JavaScript this page loaded, so every comparison between your answer and the key happens in this tab. Nothing about your level is transmitted, which is also why a reload loses a finished walk and why the copy button carries the run seed — the seed is the only thing that can rebuild the order the items came in.