PTE Core · Part 1 · Listening + Speaking
PTE Core Repeat Sentence
Ten to twelve short sentences, each heard once and never shown on screen. You say each one back exactly as you heard it. It is the task that scores Listening and Speaking together, and the one where a few seconds of focus changes the most.
- Questions in the test
- 10–12
- Time
- 3–9 s prompt, 15 s to respond
- Scoring
- Partial credit
- Traits scored
- Content, Pronunciation, Oral fluency
260 real Repeat Sentence questions inside
A glimpse of the bank. The full prompt, audio and scoring open after you sign in.
- 1We will change the classroom because this one is too small.
- 2The professor will talk about the summary in the lecture.
- 3Distance learning has become far more popular these days.
What the task asks
A sentence plays through your headset. It is between three and nine seconds long, which in practice means roughly eight to sixteen words. Nothing is written on screen except the instruction. As soon as the audio finishes the recording status changes and the microphone is live; there is no separate countdown to prepare. You have fifteen seconds to say the sentence back, and as with every speaking task, three seconds of silence closes the microphone and submits whatever you have said. Each sentence plays once. There is no replay button and no way to go back.
The sentences sound like everyday working English rather than textbook lines: a note about a delivery, a line from a safety briefing, a comment a colleague might make in a meeting. They are spoken by a range of voices and accents at a natural pace, so the task is testing whether you can hold a complete idea in your head for a few seconds and reproduce it, not whether you can decode slow, careful speech.
How it is scored
Each response is scored on three traits. Content feeds your Listening score and your Speaking score; pronunciation and oral fluency feed Speaking only. Because there are ten to twelve of these items, Repeat Sentence carries more weight in the final Speaking number than its short length suggests.
| Trait | What is measured | Scale |
|---|---|---|
| Content | How much of the sentence you reproduced, in order. All words in the correct sequence scores 3; at least half in sequence scores 2; less than half scores 1; almost nothing scores 0. | 0 to 3 |
| Pronunciation | Whether the sounds, word stress and rhythm are easy for a listener to follow. | 0 to 5 |
| Oral fluency | Smoothness of delivery: hesitations, repetitions, false starts and long pauses lower it. | 0 to 5 |
Two details are worth holding on to. First, the content trait cares about words in sequence, so saying the first eight words cleanly and stopping can still earn 2 of 3, while scrambling the order of a complete sentence can cost you. Second, hesitations and fillers are ignored for content. An um in the middle does not count as a wrong word. It does count against fluency, though, so it is never free.
How our practice scorer grades it
When you record a Repeat Sentence here, the recording is transcribed and matched word by word against the prompt. The matcher uses a short look-ahead window, so if you skip a word the rest of the sentence still lines up, and for this task it also accepts out-of-order matches and listed alternate words, because a sentence said with one phrase swapped is still mostly right by Pearson's rule. You see the prompt with every word you hit and every word you missed highlighted. Fluency is measured from the audio itself, looking for mid-phrase pauses and erratic pitch changes, and your words per minute are counted only while you are actually speaking, so a slow start does not inflate or deflate the figure. Half-finished attempts and recordings where the microphone failed are marked partial or void and never pull your averages down. The result is shown on the 10 to 90 scale with the matching CLB level.
Treat the number as a practice estimate. It is built to rank your attempts the way the real scorer would, so a rising trend here is real progress, but no practice tool sees your actual test.
A worked example
The library extends its opening hours during the final two weeks of every term.
Fourteen words, about five seconds of audio. Nobody holds fourteen separate words in short-term memory; what you hold is the meaning, and the words come back attached to it. Hear it as three chunks: the library extends its opening hours (what happens), during the final two weeks (when), of every term (how often). The speaker will put a slight stress on extends, final and term, and those stressed words are the hooks. If you catch the hooks and the shape of the sentence, the small grammar words fill themselves in.
Suppose you lose the middle. Saying The library extends its opening hours ... of every term with a confident pace still gets most of the words in the right order, which is a content score of 2, with pronunciation and fluency intact. Saying The library ... um ... opening hours ... two weeks, every term gets fewer words, in a broken order, and the hesitations drag fluency down as well. The lesson is that a clean partial answer beats a complete but stuttering one.
Five habits that raise the score
- Listen for the meaning, not the words. Picture what the sentence describes while it plays. Meaning is sticky; strings of words are not.
- Hear it in chunks. Most sentences at this length break into two or three natural phrases. Learn to feel where the speaker breathes and store the phrases, not the syllables.
- Start the instant the audio ends. The microphone is already open. A two-second think costs fluency and risks the sentence slipping away.
- Keep the rhythm even if you lose a word. Say what you have at a steady pace. Never stop, apologise or restart; the gap is counted either way, and a smooth delivery protects two of the three traits.
- Copy the speaker's stress and intonation. Repeating the melody of the sentence, not just its words, helps your memory and is exactly what the pronunciation trait rewards.
The mistakes that cost the most
- Stopping after the first half. People who lose the end of the sentence often freeze. Say the half you have and finish the thought with a falling tone; you still collect partial content.
- Paraphrasing. Opening times for opening hours, each for every. Synonyms are not the prompt. The content rule counts the words you were given.
- Whispering along with the audio. Mouthing the words while the sentence plays splits your attention and you lose the second half. Listen in silence, then speak.
- Filling silence with um. Fillers do not cost content, but they are hesitations, and hesitations are what the fluency trait measures.
- Reaching for the noteboard. A five-second sentence is gone before you have written three words. Note-taking belongs in Summarize Spoken Text, not here.
Quick answers
How many Repeat Sentence questions are in PTE Core?
Ten to twelve, all in Part 1, straight after Read Aloud. Together they are the largest single block of speaking items in the test.
Does Repeat Sentence count toward Listening?
Yes. The content trait contributes to Listening as well as Speaking, which is why a weak Repeat Sentence shows up in two scores. Our CLB converter shows how each of those scores maps to a benchmark level.
What if I only remember the end of the sentence?
Say the part you remember. Words in the correct sequence count even when they come from the end, and a confident partial answer still earns pronunciation and fluency points.
Can I hear the sentence again?
No. Every audio prompt in PTE Core plays once. Practising with single plays, as the rest of the test demands, is the only way to make the single play feel normal.