IELTS Speaking Band Descriptors Explained: What Examiners Actually Score

What are the four IELTS speaking band descriptors?

Every IELTS examiner uses the same public document: a grid of four criteria running down the side and nine bands, from 0 to 9, running across the top. Those four criteria are Fluency and Coherence, Lexical Resource, Grammatical Range and Accuracy, and Pronunciation. Each one is scored separately, then the four are averaged to produce your speaking band.

The IELTS speaking band descriptors explained this way matters because it changes where you spend your effort. A learner who assumes the score is a single impression of "how good my English sounds" will practise everything at once and improve nothing. A learner who knows the four separate scales can find the one that is holding them down.

The full descriptors are published by the IELTS official band descriptor document, and they are worth reading in full. The language is dense, though, so the rest of this article translates each scale into what it actually means in the test room.

IELTS speaking band descriptors explained: fluency and coherence is not just speed

The most common misunderstanding about IELTS speaking is that fluency means talking fast. It does not. Fluency is the ability to keep speaking at a natural pace without long pauses or false starts, while coherence is whether your ideas connect in a way the examiner can follow. Speed is only a symptom of fluency, not fluency itself. In Tutoro, the conversation tutor keeps you talking through a full answer and then shows you where you stalled, so the difference between fluency and speed becomes something you can hear rather than guess at.

At band 6, a candidate speaks at length but with noticeable hesitation and repetition, often because they are searching for a word. At band 7, the candidate speaks at length without much effort or loss of coherence, and uses a range of connectives flexibly. At band 8, the speech flows with only occasional, content-related hesitation, and ideas are developed coherently and appropriately.

Notice the word "develop". Coherence is not just linking words like "however" and "moreover". It is whether each idea is extended logically before you move on. A band 6 candidate lists points; a band 7 candidate explains them. That distinction is cheap to practise and expensive to ignore.

Lexical resource: what counts as precise vocabulary

Lexical Resource scores how wide and how precise your vocabulary is, plus how well you paraphrase when you cannot find the exact word. The band descriptors use the phrase "less common and idiomatic vocabulary" for the higher bands, which sends many learners off memorising obscure word lists. That is a misreading.

What the descriptor actually rewards is precision. A band 7 speaker can discuss a range of topics and use some less common words and idiomatic language, with an awareness of style and collocation. Collocation is the key term: knowing which words naturally go together. Saying "make a decision" rather than "do a decision" is collocation, and it counts more than knowing a rare synonym.

At band 6, vocabulary is adequate for the topic but may be repetitive or imprecise, with errors in word choice. At band 8, the speaker uses vocabulary flexibly and conveys precise meaning, with only occasional inaccuracies. The gap between 6 and 8 is not more words; it is the same words used with more control.

This is where targeted practice pays off. Rather than learning fifty new words a week, a candidate improves faster by taking the words they already half-know and using them in full sentences until the collocations become automatic.

Grammatical range: why mistakes cost less than you think

The full name of this criterion is Grammatical Range and Accuracy, and the order of those two words is deliberate. Range comes first because the descriptors reward a candidate who attempts complex structures even when they make errors, more than a candidate who stays safe with simple sentences and makes none.

At band 6, the speaker uses a mix of simple and complex structures, but errors are frequent and may cause some difficulty for the listener. At band 7, the speaker produces a range of complex structures with frequent error-free sentences, though some grammatical mistakes persist. The phrase "error-free sentences" is the target: you do not need a perfect test, you need stretches of accuracy.

This reframes the common advice to "avoid mistakes". Avoiding mistakes by only using simple sentences caps your score at band 6, because you are not demonstrating range. The higher-band strategy is to attempt subordinate clauses, conditionals and relative clauses, accept that some will be wrong, and make sure enough are right that the examiner hears the pattern.

What separates a band 6 from a band 7 in grammar is often just one habit: the band 7 candidate self-corrects. When they hear themselves make an error, they fix it, and the descriptor explicitly rewards this.

Pronunciation: the criterion most learners ignore

Pronunciation is worth a full quarter of your speaking score, yet most self-study plans give it almost no time, because it is hard to practise alone. The descriptor does not require a native accent. It requires that your speech be easily understood throughout, with control over individual sounds, word stress, sentence stress and intonation.

At band 6, the candidate uses a range of pronunciation features but control is variable, and mispronunciations occasionally reduce clarity. At band 7, all the features are used effectively and the speaker is easy to understand, with an accent that may be noticeable but does not interfere. At band 8, the features are used skillfully and are easy to understand throughout.

The word "features" is doing the work here. It is not one thing. It is individual sounds, the stress inside words, the rhythm across a sentence, and the rise and fall of intonation. A learner who only fixes individual sounds, like the difference between "ship" and "sheep", is improving one feature and leaving three untouched.

This is the criterion where feedback matters most, because you cannot hear your own intonation the way an examiner will. Practising against a source that scores your speech word by word gives you the feedback loop that silent study cannot.

That is exactly the gap an AI English tutor is built to close. The speaking tutors listen to your replies and score each word as good, fair or poor as you say it, so the pronunciation feature you are weakest on shows up immediately rather than after the test.

How the four scores become one band

Your speaking score is the arithmetic mean of the four criterion scores, rounded to the nearest half band. If you score 7 for Fluency and Coherence, 6 for Lexical Resource, 6 for Grammatical Range and Accuracy, and 7 for Pronunciation, the average is 6.5, and your speaking band is 6.5.

This averaging has a consequence most candidates miss: a single weak criterion drags the others down, but a single strong one rarely lifts them up. Scoring 8 in Fluency and Coherence while scoring 6 in Grammar gives you 6.5 overall, the same as a candidate who scored a steady 6.5 everywhere. Balance beats brilliance.

It also means the fastest route to a higher band is usually the lowest criterion, not the one you already enjoy. A candidate who loves speaking but hates grammar study will get more points from lifting grammar from 6 to 7 than from pushing fluency from 7 to 8, because the floor is what holds the average down.

The IELTS calculation method confirms that the four scores are averaged, not weighted differently, which is why examiners score each criterion one at a time rather than forming one overall judgement.

How long should you practise before the test?

There is no honest one-size answer, because the size of the gap matters. A candidate already at a solid band 7 in three criteria who needs to lift pronunciation can see movement in four to six weeks of focused daily work. A candidate at band 5 overall needs months, because every criterion has to move at once.

What is consistent is the shape of effective practice. It targets one criterion at a time, it produces spoken output rather than silent reading, and it gives you feedback on the same features the examiner scores. Watching English videos builds input, which helps, but the speaking test is an output test, and output only improves by producing it.

Speaking practice with an AI tutor works well for this because it is always available and never gets tired of repeating a question. You can run the same Part 2 topic three times in a row, each time focusing on a different criterion: once on coherence, once on range, once on pronunciation. A human teacher will rarely sit through that repetition, but it is exactly what moves a band.

An AI English teacher gives you that repetition without judgement. You can ask for the same speaking prompt again, hear your own errors scored as you make them, and watch the weak criterion move before you sit the real test.

Tutoro runs speaking practice inside a conversation with a 3D tutor who listens and reacts while you talk, so the repetition a band lift requires is always available. Pronunciation is scored word by word as good, fair or poor, which makes the one criterion most learners cannot practise alone finally measurable.

It is not a substitute for a full mock test with a human examiner, and if you need a scored simulation of the exact test format, you should still book one, but for the daily output practice that moves a band it removes the usual barriers of cost and scheduling.

What should you do differently after reading this?

Start by finding your floor. Record yourself answering one Part 1 and one Part 2 question, then score each criterion honestly against the public descriptors. Most candidates are surprised by where their weakness actually is, and it is rarely the one they assumed.

Then spend the next two weeks on that single criterion. If it is coherence, practise developing every answer to three sentences before you stop. If it is grammar, force yourself to attempt one subordinate clause per answer and self-correct when it goes wrong. If it is pronunciation, read aloud and score yourself against a recording.

The IELTS speaking band descriptors explained properly are not a mystery to decode. They are a checklist, and the candidate who treats them as one, finds the lowest box, and works on it deliberately, will score higher than the candidate who practises everything vaguely and hopes.