Free workplace English test

Eight questions, an advisory 9-minute clock, scored the moment you press the button. No account, no email field, no paywall. Everything — the items, the key, the explanations — is already in this page, and the scoring happens in your browser, so nothing you do here is sent to AssessAll or stored anywhere.

Two things make this different from the other free English tests you will find today. The first is what it tests. The lane is full of vocabulary and grammar quizzes; these eight items are the constructions that actually cost a distributed team a working day — a sentence that estimates when the reader heard a commitment, a date written 09/10, a blocker with no owner in it, a pronoun with three possible antecedents in a ticket handed over while you slept, and next Friday.

The second is what it refuses to do. It will not tell you that you are B2. Nothing that takes eight questions and nine minutes can tell you that honestly, and the section below shows you the arithmetic, plus the sentence about it from the organisation with the most authority to write one.

The test: 8 questions, 9 minutes

Answer as many as you like and press Score my answers. Unanswered items count as wrong. You can restart as often as you want, and every item comes back with the reasoning for all four options, not only the one you chose.

Why this page will not tell you that you are B2

Start with the strongest available authority, which is not us. IELTS publishes a page on how its band scores relate to the Common European Framework, and it says this:

“There is not a one-to-one correspondence between IELTS scores and CEFR levels.”

The same page describes band 8 as borderlinefor C2, and says of band 5 that “it is not clear whether band 5 is B1 or B2”. Hold that next to what it is describing: a two-to-three hour examination of all four skills, sat under invigilation, with decades of calibration and a research programme behind it. If the relationship between that and the CEFR is a range with borderline cases, then a browser quiz of 8 multiple-choice items that returns a single letter is not performing a smaller version of the same operation. It is performing a different operation and borrowing the name.

Now the arithmetic on this end. Every score out of 8 carries a confidence interval, and on 8 items those intervals overlap so heavily that most scores cannot be told apart from most other scores. The table is computed with the Wilson score method rather than the textbook formula, which misbehaves badly at this sample size:

Someone scoring 5 out of 8 has a true rate somewhere between 30.6% and 86.3%. Someone scoring 7 out of 8 has a true rate somewhere between 52.9% and 97.8%. Those ranges overlap across more than thirty percentage points. Now ask the level question on top of that: the CEFR has 6 levels, so assigning one means cutting a scale into 6 pieces using a measurement whose own uncertainty is wider than several of the pieces put together. That is not a hard problem. It is an unanswerable one at this length.

There is a second way to see it. The Spearman–Brown prophecy formula predicts what happens to reliability when a test changes length. A full-length four-skills form of 60 items with a reliability of 0.85 — both figures illustrative of the class rather than measurements of any AssessAll product, and stated as such — drops at 8 items to 0.43. And that assumes the 8 items are a parallel sample of the full form, which flatters the short version: a hand-picked 8 is usually less internally consistent than a random 8, not more.

What it would actually take to give you a CEFR level

Four things, and a short test can supply none of them:

1. A sample of all four skills. The CEFR describes what a person can do in listening, reading, speaking and writing. A reading-and-usage quiz observes one of the four, so any level it produced would be an extrapolation from a quarter of the evidence.
2. Items calibrated against the descriptors. Not written to feel like a level, but empirically placed on a difficulty scale with the level descriptors anchored on the same scale.
3. A standard-setting study. A documented procedure, run with trained judges, that decides where on the scale one level ends and the next begins — and reports how much the judges disagreed.
4. A published standard error. Without it, nobody can say how often the instrument puts a person one level away from where they belong, which is the only figure that makes a level claim checkable.

A reasonable question at this point is why a test that will not assign a level is worth nine minutes. Because the level was never the useful part. What eight items can give you is a diagnosis — which of the eight constructions you are getting wrong, with the reasoning for all four options printed whether you chose them or not. That does not depend on the score being reliable, and it is the part you can act on tomorrow morning.

What AssessAll actually prints, and what it refuses to

AssessAll is an assessment platform rather than a test-prep site, and it does print CEFR bands on its English reports — so the argument above has to be applied to our own product before it is worth anything. Here is what the code does. English scores are reported on a 10–90 scale with 6 bands mapped onto it, and the boundaries are documented in the source as the launch calibration — design priors a calibration study can move — rather than empirically derived cut scores:

Two consequences follow directly from that table and neither is flattering. A band is up to 16 points wide, so two people both described as B2 can be 13 points apart. And a single point of movement changes the letter: a composite of 66 prints B2 and a composite of 67 prints C1, on a difference no reader of the report can see. An employer screening on the letter alone is treating a 16-point range as a point.

Where the platform is asked for alignment to a named exam framework it goes further and refuses to print a point estimate at all. The module that produces those figures returns a range, by design, and every surface that displays one is required to display the disclaimer beside it: an estimated range from independent practice material, not affiliated with, endorsed by, or a substitute for the exam it is estimating. That is the same refusal this page makes, enforced in code rather than in copy — and it is the reason we can make the argument above without hypocrisy.

What AssessAll does not publish is a reliability coefficient, a norm table, a benchmark or a percentile for any of its own instruments. The platform holds an internal research-readiness gate — at least 400 completed sittings on the single assessment being reported, from at least ten distinct organisations, with no organisation contributing more than a quarter of the sample — and no instrument has yet cleared it. A page arguing that eight items is not enough evidence cannot itself publish an under-evidenced number, so it does not.

When to use something else instead

Three alternatives, opened and read on 14 September 2026, and the case for choosing each over this page:

• An exam an employer or visa route names — IELTS, TOEFL, Versant, Cambridge. If someone has named an instrument, nothing else answers the request, and no free test changes that. One rule worth knowing before you book: IELTS One Skill Retake lets you resit a single section — Listening, Reading, Writing or Speaking — but only within 60 days of the original test, only once per full test, and only where the original was taken on computer at a participating centre. Paper tests are not eligible. That 60-day window is the reason to diagnose before you book rather than after.
• EF SET, for a free full-length result you can show someone. Fifty minutes for two skills or ninety for four, CEFR-aligned, free, and the result is intended to be usable as a language certification on a CV or a LinkedIn profile. If what you need is a level to put in front of an employer rather than a diagnosis to act on, take theirs, not this. Worth noting alongside it: their score page asserts CEFR alignment and points to a research section, but publishes no reliability figure or standard error on the page itself — which is the question to put to any test that hands you a letter.
• A vocabulary-only level checker, for a four-minute estimate. The better ones are explicit about their scope. One we read runs 24 adaptive four-choice vocabulary items in four to six minutes, needs no signup, and states on its own page both that it is not an official language certificate and that it does not assess four-skills proficiency. That honesty is the exception in this lane rather than the rule — several results on the same search advertise a certificate in the page title — and it is the thing to look for. None of them, including the honest ones, prints a margin of error next to the level it gives you.

Stated as facts about published pages read on the date above, not as judgments of instrument quality. Where a test does something well, the row says so.

What this page is not

• This page does not return a CEFR level, and that is deliberate rather than a limitation we intend to remove. Eight multiple-choice items cannot place anyone on a six-level scale, and the section below shows the arithmetic rather than asserting it.
• It tests reading and usage only. Speaking, listening and writing production — three of the four skills any full English test covers — are entirely absent, so a strong score here says nothing about how you sound on a client call.
• There is no certificate and nothing to download at the end. Several free tests in this lane issue one; a certificate generated from eight items is a decoration, and putting one on a CV is a risk rather than an advantage.
• It is not a sample of an AssessAll product. Every item was written for this page. Nothing here is drawn from the live item bank or from any form a candidate might sit.
• It cannot predict a result on IELTS, TOEFL, Versant, or any employer's English screen. Where an employer or a visa route names an instrument, the only thing that answers it is that instrument.
• It is not a general English proficiency test. It targets a narrow, high-consequence slice — the constructions that break async work across time zones — and a person can be strong on this slice and weak elsewhere, or the reverse.

Common questions about workplace English tests

Read next

• Know your IELTS band before you book

  The corridor version of the same argument: what a band is worth, what One Skill Retake changes, and why booking before you have a diagnosis is the expensive order.

• CEFR levels explained

  What A1 to C2 actually describe, and why a level is a range of behaviour rather than a score.

• How to verify remote candidate skills

  The employer's side of this page: what a US or UK buyer can and cannot establish about a remote candidate, and what the evidence looks like.

• Free numerical reasoning practice test

  The sibling teaser, and the same confidence-interval argument applied to an aptitude score.

• Voice & service suitability report, annotated

  What a platform-mapped CEFR level looks like on a real report, panel by panel, including the one-point movement that changes the letter.

Sources

• IELTS, “IELTS and the CEFR” (read 14 September 2026) — The exam board's own statement that there is no one-to-one correspondence between its band scores and CEFR levels, and that bands 5 and 8 are borderline.
• IELTS, “One Skill Retake” (read 14 September 2026) — The 60-day window, once per full test, computer-delivered only — the rule that decides whether a diagnosis before booking is worth anything.
• Wilson, E. B. (1927), Journal of the American Statistical Association 22(158), 209–212 — The score interval used for the confidence band on the result.
• Agresti, A. & Coull, B. (1998), The American Statistician 52(2), 119–126 — Why the textbook Wald interval is the wrong tool at n = 8.
• EF SET, score and certification page (read 14 September 2026) — The free full-length comparator: two skills in 50 minutes or four in 90, CEFR-aligned, with a result usable as a certification on a CV.