A lecture of up to 90 seconds, retold in your own words after ten seconds of preparation. The task that punishes note-taking systems.
“You will hear a lecture. After listening, in 10 seconds, please speak into the microphone and retell what you have just heard in your own words.”
Ten seconds is not enough to read a paragraph you wrote. Capture five or six single words — the topic and the main points — and nothing else.
Eyes on notes produces the pauses the fluency trait penalises. Look once during the preparation, then speak from memory.
Cover the whole lecture thinly rather than the first third thoroughly. Content credit is spread across the lecture, not weighted to its beginning.
It is a fixed cost of roughly three seconds out of forty and adds no content. Name the topic in your first clause instead.
The ten-second gap is the binding constraint, and it means the note-taking systems that work for university lectures actively hurt here. Design your notes for a ten-second read.
Six texts of up to 60 words, read aloud from the screen. The easiest task to sound fine on and still score badly.
Ten sentences, heard once, repeated exactly. Scored on what you reproduce, not on whether you understood it.
A chart, map, graph or process diagram, described from scratch. The task where a rehearsed structure genuinely helps.
A short general-knowledge question, answered in a word or two. The only Speaking task scored on listening comprehension alone.
Up to 90 seconds. You then get 10 seconds to prepare and 40 seconds to speak. Three of these appear per sitting — the fewest of any Speaking task, but the longest individual items.
Yes, but only five or six single words. Ten seconds is not enough to read a paragraph you wrote, and eyes on notes during the recording produces exactly the pauses the fluency scoring penalises.
No — unlike Repeat Sentence, your own words are expected here. Naming the topic, the main claim and two supporting points covers most of the available content credit.
The whole lecture, thinly. Content credit is spread across it rather than weighted to the opening, so candidates who detail the first third and run out of time lose marks that were available.
Our PTE mock runs Re-tell Lecture on the published timings — 10s to prepare · 40s to answer · 3 per sitting — and marks your recording word by word against the prompt. The first session is free.
This guide is written from the PTE Academic test format published by Pearson, in our own words. For the official wording, see Pearson's Speaking & Writing format page. MockLive AI is an independent practice tool and is not affiliated with Pearson.