The IELTS Speaking band descriptors are published by the test owners on ielts.org as a four-page PDF that sets out nine bands against four criteria: Fluency and Coherence, Lexical Resource, Grammatical Range and Accuracy, and Pronunciation. The same descriptors cover Academic and General Training, and each criterion carries equal weighting in the Speaking score. Examiners assess speaking throughout the face-to-face interview, so the four bands reflect your speech across the whole test.
Read as a ladder, with each band listing what a speaker can do and what still limits them, the descriptors show which features repay practice first and give a fixed checklist for marking your own recordings. The paraphrases here follow the published wording closely enough to use alongside it, whilst the official PDFs remain the reference for exact phrasing.
Where are the official IELTS Speaking band descriptors published?
The official public descriptors sit on ielts.org as a four-page PDF titled Speaking Band Descriptors, headed as scoring criteria for Academic and General Training tests, so one document serves both versions. Each of the four criteria has its own column and each band its own row, running down from band 9. A companion four-page PDF, Speaking Key Assessment Criteria, is linked from the ielts.org scoring page and explains what each criterion looks for. The descriptors carry a note asking readers to visit ielts.org for updates, so the live file is the one to print.
IELTS is jointly owned by the British Council, IDP IELTS and Cambridge University Press & Assessment, a statement that appears on the key criteria document. Reading that document first usually makes the band descriptors easier to follow, because terms such as discourse markers, phonological features and paraphrase are unpacked there. Keeping both open whilst you mark practice recordings saves repeated searching.
What are the four Speaking criteria and how are they weighted?
Examiners award a score for Fluency and Coherence, Lexical Resource, Grammatical Range and Accuracy, and Pronunciation, and each criterion carries equal weighting, with the overall average giving the Speaking score. Pronunciation therefore counts for a quarter of the result, exactly as Fluency and Coherence does.
Fluency and Coherence covers continuity, speech rate and the logical linking of ideas into connected speech. Lexical Resource covers the variety and precision of vocabulary, including paraphrase when a word is missing. Grammatical Range and Accuracy covers the length and complexity of sentences together with how many errors occur and how far they affect meaning. Pronunciation covers the phonological features that carry meaning to a listener.
IELTS Preparation tutors
Mark
Master TutorI have thirty years in education, including Head of English at an independent school and GCSE and A-Level examiner. I teach English, Literature, Language, Creative Writing and Critical Thinking, ages four to eighteen. I start from what a student can already do and adapt my methods to them.
Enquire about this tutor
Johanna
Expert TutorI have two decades' teaching experience in English Literature, English Language and Philosophy, and hold a PhD in English Literature. I teach GCSE, A-Level, Oxbridge entrance and UCAS personal statements. My approach is person-centred, working through dialogue and ongoing assessment.
Enquire about this tutor
Owen
Expert TutorI gained an MA and MPhil at Girton College, Cambridge, and am a British Council certified counsellor. I specialise in Oxbridge preparation and university admissions, and teach A-Level English Literature, French and German, using the discursive, oral-first style of an Oxbridge supervision.
Enquire about this tutorWhat separates neighbouring bands on each criterion?
On Fluency and Coherence, the step between bands 8 and 7 concerns why a speaker pauses. At band 8 most hesitation is content related, with only occasional searching for words or grammar, whereas band 7 allows some hesitation, repetition and self-correction, often mid-sentence, that signals trouble finding language without harming coherence. At band 6 the speaker is willing to produce long turns, yet coherence may be lost at times. Band 5 relies on repetition, self-correction or slow speech to keep going, and band 4 cannot continue without noticeable pauses, with slow speech, frequent repetition and frequent self-correction.
On Lexical Resource, neighbouring bands differ in how flexibly and precisely vocabulary is used. Band 8 has a wide resource used readily across all topics, with skilful use of less common and idiomatic items despite occasional slips in word choice and collocation, whereas band 7 shows some ability with such items and an awareness of style and collocation, though inappropriacies occur. At band 6 the vocabulary is enough to discuss topics at length and meaning stays clear, with generally successful paraphrase. Band 5 has limited flexibility and mixed success with paraphrase, and band 4 conveys only basic meaning on unfamiliar topics.
On Grammatical Range and Accuracy, neighbouring bands differ in how often sentences are error free and how far complex structures hold up. Band 8 has a wide range with most sentences error free and only occasional non-systematic errors, though a few basic errors may persist, whereas band 7 has frequent error-free sentences, effective simple and complex forms, and a few basic errors that persist. Band 6 mixes short and complex forms with limited flexibility, and errors in complex structures are frequent yet rarely impede communication. At band 5 basic forms are fairly well controlled but complex structures nearly always contain errors, and band 4 has rare subordinate clauses and frequent errors.
On Pronunciation, neighbouring bands differ in how many phonological features are used and how consistently they are controlled. Band 8 uses a wide range, sustains rhythm, varies stress and intonation across long utterances despite occasional lapses, and is easily understood. Band 6 uses a range with variable control, generally appropriate chunking, stress and intonation that are not sustained, and only occasional lack of clarity. At Band 4, the range is limited, rhythm lapses frequently and mispronounced sounds make understanding require some effort. Bands 7 and 5 carry no separate wording and are defined as showing the positive features of the band below plus some, but not all, of those of the band above.
Does the examiner rate each part separately or the whole test?
The examiner assesses speaking performance throughout the test, which is a recorded face-to-face interview in three parts lasting 11 to 14 minutes in total. Part 1 is an introduction and interview of 4 to 5 minutes on familiar topics, Part 2 is a long turn of 3 to 4 minutes including one minute of preparation with a task card, and Part 3 is a discussion of 4 to 5 minutes on broader issues related to the Part 2 topic. The scoring page describes one score for each of the four criteria, so the three parts together supply the evidence for those scores.
Because assessment runs throughout, the judgement draws on everything heard from the first question to the last. Practice therefore works best when it builds consistency across short answers, a sustained turn on your own and abstract discussion. The higher descriptors refer to features sustained across long utterances or throughout, and the long turn and the discussion are where that sustaining becomes audible.
What does pronunciation mean in the Speaking band descriptors?
In the descriptors, pronunciation means the accurate and sustained use of a range of phonological features to convey meaning, and intelligibility sits at the centre. The key criteria list dividing speech into meaningful chunks, rhythm and stress timing, linking sounds in connected speech, stress and intonation that enhance meaning, individual vowel and consonant sounds and word stress, and the effort a listener needs to understand them. The final indicator is the overall effect of accent on intelligibility, and band wording follows it: at the top, accent has no or minimal effect, and at lower bands understanding requires effort.
Accent is not scored in its own right, since the descriptors ask only how far it affects understanding. Official guidance states that examiners do not judge how native-like an accent sounds, and a speaker can keep an accent whilst adjusting the sounds, stress and rhythm that affect clarity. Practice should target features listeners rely on, such as stressing the right syllable and grouping words into phrases, instead of imitating a particular accent.
How can you record and self-assess practice answers against each criterion?
Recording practice answers on a phone and listening back later gives the most direct check against the four criteria, since the live test is itself recorded. Record in the shape of the interview: short answers to questions on familiar topics, a long turn after one minute of preparation, and a longer discussion of abstract questions on the same topic. Listen once for meaning without stopping, then replay with the descriptors open.
For Fluency and Coherence, mark each pause, repetition and self-correction on a transcript, and note whether it came from planning the next idea or searching for a basic word. Check whether sentences follow a logical order and whether connectives and discourse markers vary. For Lexical Resource, underline less common words and collocations, mark any paraphrase used when a word was missing, and note vocabulary that did not suit the style of the question.
For Grammatical Range and Accuracy, transcribe a minute of speech, count the error-free sentences, and mark subordinate clauses, perfect and continuous verb forms, modals and passives. Note which errors would confuse a listener, because the key criteria weigh the effect of an error on meaning alongside how many occur. For Pronunciation, listen without the transcript and mark any word you struggle to catch, then check chunking, word stress and whether stress and intonation shift meaning. A friend who has not seen the transcript can list the words they missed.
After rating each criterion against the descriptor wording, read the row above the one that seems to fit and list the phrases not yet met. Repeating the exercise across several recordings on different topics shows which features recur, and judging patterns across recordings is more reliable than judging one. Those recurring features set the next practice target.
In what order should you practise for each criterion?
A workable order starts with the documents and ends with full-length recordings, because the descriptors reward features that appear throughout the test. Short, frequent recordings tend to reveal more than occasional long sessions, and keeping dated notes against each criterion shows which features have shifted.
- Read the Speaking Key Assessment Criteria and the band descriptors side by side until the terms are familiar.
- Build continuity with timed long turns, since continuous speech gives the other three criteria something to judge.
- Add connectives and varied structures, including subordinate clauses, to those long turns.
- Widen vocabulary by topic and practise paraphrasing words you cannot recall.
- Work on chunking, word stress and intonation by listening back, then repeat the full interview shape and re-mark it.
Related guides
- What is IELTS? Format, versions and scoring
- IELTS band scores explained
- IELTS Listening and Reading raw scores to bands
- IELTS Writing band descriptors
- IELTS Writing Task 1 (Academic)
- Task 1 line graph model answer
- Task 1 bar chart model answer
English tuition for adults with Taylor Tuition
Adults who come to us for English usually know the situation they need it for, whether that is a job, a test or daily life in the UK, so lessons start with a level check against the CEFR scale and are then built around the tasks the learner actually faces.
English for adults is taught one-to-one or in small groups, at every level from beginner (CEFR A1) to proficient (C2), online or in person in central London. Fees start from £50 per 50-minute session and depend on tutor level, there is no registration fee and the initial consultation is free. Employers can book programmes for staff and are invoiced directly.
Read more about our IELTS preparation tutors, or send an enquiry and we will come back with tutor recommendations.
Frequently asked questions
- Are the Speaking band descriptors the same for Academic and General Training?
- The published descriptors are headed as scoring criteria for Academic and General Training tests, so one document covers both. Candidates preparing for either version can mark practice against the same four criteria and the same nine bands. The ielts.org file also carries a note asking readers to check the site for updates, so downloading it again before a final round of practice is sensible.
- Is IELTS Speaking marked by an examiner or by a computer?
- The ielts.org format page describes the Speaking test as a face-to-face interview between the test taker and an examiner, and says certificated IELTS examiners assess performance throughout. The test is recorded. Practice recordings copy those conditions most closely when a real listener plays the examiner and asks the questions, because speaking to a person differs from speaking to a phone.
- Do idioms and long words raise the Lexical Resource band?
- The criterion asks for adequate and appropriate vocabulary: correct labelling of things and ideas, a suitable style, sound collocation and a clear attitude to the content. Idiomatic items appear in the upper band wording as skilful when they fit, and inappropriacies are named as a limit. A plain word used precisely therefore serves the criterion better than an idiom used awkwardly.
- Does speaking faster improve Fluency and Coherence?
- The key criteria say speech rate should ideally not be too slow, and they also value flow that false starts, backtracking and word searches do not interrupt. Pronunciation wording at band 6 notes that a rapid speech rate can affect rhythm. A steady pace with continuous, well-chunked speech meets both points, so timing practice with a recording is more useful than simply speeding up.
- Can memorised answers help in the Speaking test?
- Memorised language appears in the descriptors only at the lowest bands, where grammar errors are numerous except in apparently memorised utterances and vocabulary is limited to isolated words or memorised utterances. Higher bands ask for flexible use of structures, topic development and paraphrase. Practising from varied prompts, so that you build answers as you speak, matches that wording.
Enquire about IELTS preparation
Start your enquiryHow Taylor Tuition can help
Expert, one-to-one support from specialist tutors.
Editorial Notice
We use automated research tools to help collate and maintain information across our website. While each post is reviewed before publication, some details may be incomplete or inaccurate.
Taylor Tuition
Educational Consultancy
Contributing expert insights on education, exam preparation, and effective learning strategies to help students reach their full potential.