PTE template · Framework

PTE Repeat Sentence Template: chunk, mimic, deliver for Band 79

Repeat Sentence plays you a 3 to 9 second sentence exactly once, then opens the microphone for a 15 second answer. There are 10 to 12 items per PTE Academic test, all scored on Content, Oral Fluency and Pronunciation. Content on Repeat Sentence is unusual: it is one of the few speaking tasks scored primarily on how many words you got right in the correct order. The framework below is a repeatable listen-and-deliver routine, not a memorised script (there is no script here, because the sentence changes every item).

Take a free scored mock →All PTE templates

Quick answer

The Repeat Sentence template is a 3-step routine you apply to every item: chunk the sentence you hear into 3 to 6 meaning groups as it plays, mimic the speaker's rising and falling pitch on the last word you heard, then deliver the whole sentence in one continuous take at 140 to 160 words per minute. Start speaking within 1 second of the microphone opening; do not wait for the tone (there is no tone on Repeat Sentence).

Read this first

There is no memorisable template answer on Repeat Sentence, because the sentence changes every item. What you can standardise is the LISTENING and DELIVERY routine. Do not fill your answer with stock filler ('Um, let me think, the speaker said...') because filler adds words that are not in the original sentence and Content marks fall for every substitution, omission or insertion.

The framework

How the framework works

Read the sections in order. Each one is a step of the framework, with adaptable sentence starters you fill from the actual prompt.

The 3-step routine (per item)

1. LISTEN as chunks, not words. As the audio plays, group what you hear into 3 to 6 meaning chunks. A typical 8 word sentence chunks into 3 groups; a 15 word sentence chunks into 5 or 6. Chunking is how short-term memory holds more. 2. MIMIC the prosody. Note the pitch on the LAST word you heard. Was it rising (a question) or falling (a statement)? Copy that pitch pattern when you speak. Prosody mimicry lifts Oral Fluency and Pronunciation together. 3. DELIVER in one take. Start speaking within 1 second of the microphone opening. Deliver at 140 to 160 words per minute in one continuous pass. Do not restart, do not correct, do not add a preface.

How to chunk a sentence as it plays

Sentences arrive in natural phrase groups. Your job is to hold each phrase as a single unit rather than as a list of separate words. Example sentence (12 words): 'The library will be closed on Friday due to a scheduled maintenance shutdown.' Chunk 1: 'The library will be closed on Friday' Chunk 2: 'due to a scheduled maintenance shutdown' Holding two chunks in memory is far easier than holding twelve words. The trick is to listen for verbs and prepositions, which typically mark chunk boundaries: 'will be closed', 'due to'. With practice, an 8 second audio (roughly 15 to 20 words) settles into 4 to 6 chunks that you can recall in order.

What to do in the 1 to 2 second gap before your microphone opens

1. Silently rehearse chunk 1 in your head. Just the first chunk, not the whole sentence. 2. Take one shallow breath. You are about to speak for 4 to 8 seconds; you need air. 3. As the microphone opens, deliver chunk 1 immediately. Do not wait for a tone: Repeat Sentence has NO tone, and the recording will auto-stop if you sit silent for more than 3 seconds. Starting late costs you nothing on Content, but it steals recovery room if you fumble mid-sentence.

Delivery pace and prosody

Match the speaker's pace, not a memorised tempo. Repeat Sentence speakers deliver at roughly 140 to 160 words per minute, which is also the target pace for Read Aloud. If you speak much faster than the source, words blur and Pronunciation drops. If you speak much slower, Oral Fluency drops because the answer sounds unnatural. Pitch: copy the final word's pitch direction. A question ends on a rise; a statement ends on a fall. Copying this signals to the scorer that you heard the whole sentence, not just the words. Stress: content words (nouns, main verbs, adjectives) take a fractionally longer vowel and slight pitch rise. Function words (a, the, of, and) stay flat. Do not stress every word evenly, which reads as robotic.

What Pearson scores on Repeat Sentence

Three traits, per the July 2025 Score Guide: - Content (0 to 3): every word must be spoken in the correct order. Substitutions, omissions and insertions each cost a point. A single missing word usually costs 1 mark; a scrambled clause costs more. - Oral Fluency (0 to 5): natural, connected speech at appropriate pace. Restarts, long hesitations, and word-by-word reading all cost Fluency. - Pronunciation (0 to 5): correct stress, clean articulation, recognisable vowel and consonant sounds. Accent alone does not cost points; unclear articulation does. Repeat Sentence is AI-only scored (not one of the 7 hybrid task types). Maximum raw score is 13 per item.

The most common Repeat Sentence mistakes

- Adding a preface. 'The speaker said that the library...' inserts words that are not in the original, dropping Content marks. Just repeat the sentence. - Restarting when you fumble. Every restart costs Fluency. Push through: a wrong word costs 1 Content point, whereas a restart plus a wrong word costs both. - Waiting for a tone that never plays. Repeat Sentence has NO tone. The microphone opens, the recording status shows 'Recording', and you should already be talking. - Speaking too softly. Match the volume you used in the mic check. If the AI cannot transcribe you, Content collapses. - Trying to memorise the exact wording word by word. Chunk instead: 3 to 6 chunks is the human memory sweet spot, individual words are not.

Worked examples

The framework applied

Same framework, different prompts. Each answer is filled with real content, not a memorised script.

Example 1You hear the audio once, the microphone opens after a 1 second gap, and you have 15 seconds to answer.

Audio (8 seconds, 14 words): 'The lecture on renewable energy has been moved from Tuesday to Thursday afternoon.'

Framework-filled answer

The lecture on renewable energy has been moved from Tuesday to Thursday afternoon.

Why this scores: Chunks: 'The lecture on renewable energy' + 'has been moved' + 'from Tuesday to Thursday afternoon' (3 chunks, 14 words). Stress falls on 'lecture', 'renewable energy', 'moved', 'Tuesday', 'Thursday afternoon'. Delivery pace 140 wpm gives 6 second answer time; leaves 9 seconds of the 15 second window unused, which is fine. Fluency and Content both maximal.

Example 2You hear the audio once, the microphone opens after a 1 second gap, and you have 15 seconds to answer.

Audio (7 seconds, 12 words): 'Students must submit their assignments by five pm on the deadline day.'

Framework-filled answer

Students must submit their assignments by five pm on the deadline day.

Why this scores: Chunks: 'Students must submit their assignments' + 'by five pm' + 'on the deadline day' (3 chunks, 12 words). Note that 'five pm' is a single time chunk, not two separate items. Stress on 'submit', 'assignments', 'five pm', 'deadline day'. Delivery pace 140 wpm gives roughly 5 second answer; well inside the 15 second window.

Example 3You hear the audio once, the microphone opens after a 1 second gap, and you have 15 seconds to answer.

Audio (9 seconds, 17 words): 'The research committee will announce the successful grant applicants at the end of next month.'

Framework-filled answer

The research committee will announce the successful grant applicants at the end of next month.

Why this scores: Longer sentence, 4 chunks: 'The research committee' + 'will announce the successful grant applicants' + 'at the end' + 'of next month'. Stress on 'research committee', 'announce', 'successful grant applicants', 'end of next month'. Falling pitch on 'month' signals a completed statement. Delivery pace 145 wpm gives about 7 second answer.

Frequently asked questions

What is the best PTE Repeat Sentence strategy?

A 3-step routine: chunk the sentence into 3 to 6 meaning groups as you listen, mimic the speaker's ending pitch, then deliver the sentence in one continuous take at 140 to 160 words per minute. Start speaking within 1 second of the microphone opening; do not wait for a tone (there is no tone on Repeat Sentence).

How many Repeat Sentence items are on the PTE Academic test?

10 to 12 items per test per Pearson's July 2025 Score Guide, making Repeat Sentence the most numerous speaking task in the exam. It contributes to both your Listening score and your Speaking score, so it is one of the highest-leverage items to master.

Is there a tone before I speak on Repeat Sentence?

No. Repeat Sentence has no tone. Read Aloud, Describe Image, Retell Lecture, Summarise Group Discussion and Respond to a Situation all play a short tone before the microphone opens. Repeat Sentence and Answer Short Question do not. The microphone simply opens and the recording status changes to 'Recording'.

What happens if I miss a word in Repeat Sentence?

You lose Content marks for that word. Content is scored 0 to 3, and each missing, substituted or inserted word typically costs 1 point. Never fill a gap with a guess ('um' or 'something like') because filler words are counted as insertions and cost more Content than an honest silence.

Should I add words like 'the speaker said' before repeating?

No. Every added word is an insertion that costs Content. Start with the first word of the actual sentence. If you missed the first word entirely, start with the first chunk you did hear; that costs 1 word rather than a full extra clause.

How is Repeat Sentence scored?

Three traits: Content (0 to 3), Oral Fluency (0 to 5), Pronunciation (0 to 5). Content counts word-for-word accuracy in the correct order. Fluency counts smooth continuous delivery. Pronunciation counts stress placement and clean articulation. All three are AI-only scored (Repeat Sentence is not one of the 7 hybrid task types).

How fast should I speak on Repeat Sentence?

Match the source pace, which is typically 140 to 160 words per minute. Speaking much faster than the source blurs pronunciation; much slower reads as unnatural and drops Oral Fluency. A 15 second answer window is long enough to comfortably deliver an 18 word sentence at natural pace.