English level test: what it measures, how to read the result and when you need an exam

What an English level test result actually tells you

An English level test promises a short answer to a long question: how well do you know the language. The usual answer is a code from the Common European Framework of Reference, the scale the Council of Europe built for European languages. It has six steps from A1 to C2, grouped in pairs: A1 and A2 for a basic user, B1 and B2 for an independent user, C1 and C2 for a proficient user.

The catch is that one code describes several abilities at once. Listening, reading, speaking and writing grow at different speeds, and many learners read far better than they speak. A test that reports a single letter has either measured one skill and assumed the rest, or blended them into a number that hides the gaps. Knowing which of the two happened is the first step to reading any result honestly.

What a quick online English test can and cannot measure

A free test on a website is almost always a set of multiple-choice or gap-fill questions: pick the right tense, the right preposition, the right word for the gap. It samples your grammar, vocabulary and reading fairly well, which is enough to choose a first textbook. With only a few dozen questions, though, a handful of lucky guesses or careless slips moves the result. A second attempt at the same test mostly measures your memory of its items.

What it cannot do is hear you or read what you write. Choosing the correct form out of four options is recognition; producing it while someone waits for your answer is a different and usually weaker skill. A test can tell that you know the past perfect exists. It cannot tell whether you use it, or whether anyone understands your vowels on a phone call. That is how a person who scores B1 on grammar can still freeze in a real conversation.

Checking yourself with the CEFR self-assessment grid

The Council of Europe offers a free tool for this gap: the self-assessment grid. The version on Europass, the European Union’s site for CVs and skills, splits a language into five rows. Understanding covers listening and reading, speaking is divided into spoken interaction and spoken production, and writing stands alone. Each row has a short statement for each of the six levels, and every statement begins with the words I can.

The statements are concrete. At A2, the spoken interaction row describes someone who manages very short social exchanges but cannot usually follow enough to keep a conversation going alone. At B1, you can deal with most situations on a trip abroad and join a talk on familiar topics without preparing. At B2, the listening row expects most TV news, and the writing row an essay that argues for or against a point of view.

Use the grid row by row. Mark the highest statement in each row that you could sign today without hoping, then compare the five marks. A profile such as B1 reading, B1 listening, A2 conversation and A2 writing is normal and far more useful than one letter. It shows which skill holds the others back, and that is where the next month of practice should go.

How accurate self-assessment of language skills really is

Honest self-assessment is still a guess, and research shows how rough it is. A 2023 paper in Studies in Second Language Acquisition by Paula Winke, Xiaowan Zhang and Steven Pierce sums up an earlier review by Li and Zhang of 67 studies: on average, self-ratings and measured proficiency correlated at about .45. The two point the same way but are far from a match. The link was stronger when the statements described real-life tasks, when a rubric was used and when learners had been trained to rate themselves.

The same paper reports its own study of 807 learners of Spanish, who took a computer-adaptive speaking self-assessment and a separate oral proficiency test. The results correlated at .61, and the authors judged such self-assessments appropriate for low-stakes measurement. That is the right place for the grid and for any free test: choosing a course, a book or a starting point. When a decision has real consequences, such as admission or a visa, an independent exam does the measuring.

IELTS, TOEFL iBT and Cambridge exams compared with a free test

Official exams differ from a free test in two ways: they cover all four skills, and what you say and write is scored against published criteria. IELTS, according to its own site, consists of Listening, Reading, Writing and Speaking, and the Speaking part is a face-to-face interview with an examiner lasting 11 to 14 minutes. IELTS Academic is designed for people applying to study at undergraduate or postgraduate level.

TOEFL iBT, run by ETS, also has four sections. Its speaking tasks are called Listen and Repeat and Take an Interview, one of its writing tasks is Write an Email, and ETS asks test takers to allow about two hours. Cambridge English Qualifications, from A2 Key to C2 Proficiency, give a separate score for each skill on the Cambridge English Scale, then an overall result, a grade and a CEFR level.

You need one of these when someone else has to trust your level: a university, an employer abroad, a study or migration visa. Validity rules differ. ETS says TOEFL iBT scores are valid for two years from the test date, the IELTS partners recommend treating a result as evidence for a maximum of two years, and Cambridge English says its certificates do not expire, although each university or employer decides how recent a result it accepts.

Reading an exam score as a band, not a school grade

An exam result places you in a band, and each band is a description rather than a mark. On the IELTS nine-band scale, band 6 is called competent: you use the language effectively, with some errors, slips of usage and misunderstandings. The overall score is the average of the four sections rounded to the nearest half band, so a strong skill can hide a weak one in the total. Read the section scores first.

Links to the CEFR are approximate, and the exam boards say so. IELTS writes that there is no one-to-one correspondence: its research puts the C1 threshold between bands 6.5 and 7, and it is not clear whether band 5 is B1 or B2. Since 21 January 2026, ETS reports TOEFL iBT on a scale from 1 to 6 in half points, where a 4 in any section corresponds to B2 and a 5 to C1. Scales like these are revised, so check the current page before relying on a number.

Why learners overrate or underrate their own English

People who overrate themselves usually read well. They follow articles, recognise grammar on a test and assume that speaking will come by itself. The grid catches this quickly: the B1 line about joining a conversation without preparation is hard to sign if every sentence has to be assembled in your head first.

Underrating is just as common and harder to notice. Learners measure themselves against native speakers or films full of slang, while the B2 listening line asks for most TV news and the majority of films in standard dialect, not everything. Others go silent from nerves and judge their level by the worst five minutes of the week. The cure for both is to judge yourself by concrete tasks you did recently, such as a call, an email or a meeting.

Your first week after an English level test

Give the weakest row of your grid profile short daily practice this week. If speaking lags, record a two-minute voice note about your work or your plans, then listen and write down the words you were missing. If writing lags, write a short email you might really send. If listening lags, retell one short clip a day aloud. Retake the english level test only after several weeks, on a different one if you can, so that you measure learning rather than memory.

Practice alone has limits: nobody points out the mistake you keep repeating, and words learnt once fade without review. A structured course adds a plan and that feedback. VITON13 SCHOOL runs English in 113 Days, an online course for Russian speakers in which every explanation is in Russian, so it suits readers who read Russian. Its 113 levels, one a day, follow the CEFR scale: the core of A1, A2, B1 and B2 and a bridge to C1. It does not promise C1 or native-like speech, and IELTS and TOEFL iBT need their own preparation.

Each level brings 13 new words with sound, transcription and an example, a grammar rule with the typical mistake of Russian speakers, a short text with translation and three questions, dictation, phrase building, your own short text with a model answer and a mini-game. Words return through FSRS spaced repetition, and where the browser can recognise speech, the course checks your pronunciation. Its free placement test shows where to start, and the first three levels are open even without signing in.