How to use this bank
This is a free PTE Speaking practice bank with 40 tasks covering all seven speaking task types in PTE Academic Part 1: Read Aloud, Repeat Sentence, Describe Image, Retell Lecture, Answer Short Question, Summarize Group Discussion and Respond to a Situation. Every task shows the timer Pearson publishes for it, the traits it is scored on, a model answer written by TalkDrill (hidden until you click) and a note on what the scorer rewards. PTE Core uses five of the seven types, so Core candidates can use most of the bank as it stands. Record your answer first, then open the model.
All 40 tasks and models were written by TalkDrill for this page. They are not PTE items, not real candidates' answers, and not scored by Pearson's engine. Chart and table data are invented for practice.
- Read Aloud: 8 passages (the test has 6 to 7, per Pearson's Score Guide).
- Repeat Sentence: 8 sentences (the test has 10 to 12).
- Describe Image: 7 images, described in words until the drawings are ready (5 to 6 in PTE Academic, 3 to 4 in PTE Core).
- Retell Lecture: 4 lectures (2 to 3 in the test; Academic only).
- Answer Short Question: 5 questions (5 to 6 in the test; scored for Listening).
- Summarize Group Discussion: 4 discussions (2 to 3 in the test; Academic only).
- Respond to a Situation: 4 situations (2 to 3 in PTE Academic; 2 to 4 in PTE Core, with a longer preparation time).
Each task type has its own guide that explains the scoring and the method in depth: Read Aloud, Repeat Sentence (with Answer Short Question), Describe Image, Retell Lecture and Summarize Group Discussion, and PTE Core Speaking and Respond to a Situation. Read the guide once, then come back here to practise. Everything else about the test is on the PTE Speaking hub.
The timer for every task
These timings are from Pearson's PTE Academic and PTE Core Speaking and Writing format pages, and the item counts are from the PTE Academic Score Guide and the PTE Core Score Guide (January 2026), all checked on 21 October 2026. Part 1 of PTE Academic runs about 76 to 84 minutes and has nine question types, two of them writing. Part 1 of PTE Core runs 50 to 65 minutes and has seven, two of them writing. Set a phone timer to these numbers.
| Task | Items in PTE Academic | Before you speak | Speaking time | Items in PTE Core | Traits scored |
|---|---|---|---|---|---|
| Personal Introduction | 1, not scored | 25 seconds to read and prepare | 30 seconds | 1, not scored | None |
| Read Aloud | 6 to 7 | 30 to 40 seconds | Varies with the length of the text (up to 60 words) | 6 to 7 | Content, Pronunciation, Oral Fluency |
| Repeat Sentence | 10 to 12 | Audio of 3 to 9 seconds; no preparation | 15 seconds | 10 to 12 | Content 0 to 3, Pronunciation, Oral Fluency |
| Describe Image | 5 to 6 | 25 seconds | 40 seconds | 3 to 4 | Content 0 to 6 (0 to 5 in Core), Pronunciation, Oral Fluency |
| Retell Lecture | 2 to 3 | Audio up to 90 seconds, then 10 seconds | 40 seconds | Not in Core | Content 0 to 6, Pronunciation, Oral Fluency |
| Answer Short Question | 5 to 6 | Audio of 3 to 9 seconds; no tone | 10 seconds | 5 to 6 | Vocabulary, 1 or 0 (Listening) |
| Summarize Group Discussion | 2 to 3 | Audio up to 3 minutes, then 10 seconds | 2 minutes | Not in Core | Content 0 to 6, Pronunciation, Oral Fluency |
| Respond to a Situation | 2 to 3 | Academic: 10 seconds. Core: 20 seconds | 40 seconds | 2 to 4 | Academic: Content, Pronunciation, Oral Fluency. Core: Appropriacy 0 to 3, Pronunciation, Fluency |
Two points from the Score Guide change how you should practise. First, in PTE Academic, Describe Image, Retell Lecture, Summarize Group Discussion and Respond to a Situation all get a human expert review of Content before the score is final; everything else in speaking is scored by the AI. The PTE Core Score Guide does not list task-level human reviews. Second, Pronunciation and Oral Fluency use the same 0 to 5 scale wherever they appear, so the delivery habits you build on Read Aloud carry into every other task.
Recording yourself, and the three-second rule
The real test is a microphone and a screen. Practise the same way: your phone's voice recorder, a visible timer, one take per task. Keep the phone a hand's width from your mouth and speak at your normal volume. The Score Guide says quiet or unclear responses are the "anomalies" its machine cannot reliably score.
The rule to learn first is in Pearson's format page: "If you stop for three seconds whilst speaking, the recording status will change to 'completed' even if you haven't finished, and you won't be able to continue answering." The page names Read Aloud, Repeat Sentence, Describe Image and Retell Lecture, as well as Answer Short Question, where waiting more than three seconds to start also ends the recording. It does not mention the two newer tasks either way, so we practise as if the rule applies to every task. In practice: never stop to fix a word, never stop to think, and never say "sorry". Keep going.
For Read Aloud, TalkDrill's read-aloud practice marks your recording against the text word by word, for accuracy, fluency and completeness, so you can see exactly which words broke. It is not Pearson's engine and gives no PTE score. For Respond to a Situation, the AI role-play practice lets you rehearse everyday situations out loud with an AI partner. For every other task, your phone and the self-check sheet below are enough.
The self-check sheet
The middle column is Pearson's own question for each trait, from its format pages. The right-hand column is TalkDrill's checklist for checking a recording at home. Listen once for each row.
| Trait | Pearson's question | Check your recording |
|---|---|---|
| Content (Read Aloud, Repeat Sentence) | Does your response include all the words in the text, and only these words? | Count words changed, missed or added. For Repeat Sentence, were at least half in the right order? |
| Content (Describe Image, Retell Lecture, Summarize Group Discussion, Respond to a Situation) | Does your response accurately and thoroughly describe the image, or retell the lecture or discussion? | Did I cover the main points? Did I say how they relate? Did I end with a conclusion or implication? Were the words my own? |
| Oral Fluency | Does your response demonstrate a smooth, effortless and natural rate of speech? | Count hesitations, repetitions and false starts. Was any silence close to three seconds? |
| Pronunciation | Does your response demonstrate your ability to produce speech sounds in a similar way to most regular speakers of the language? | Which words might a stranger mishear? Is the same sound wrong every time it appears? |
| Appropriacy (PTE Core, Respond to a Situation) | Is the language used in your response relevant to and appropriate to the situation in the prompt? | Did I use the right register for this person? Was I polite? Did I deal with every part of the situation? |
| Vocabulary (Answer Short Question) | Correct or incorrect, based on how appropriate the words are | Right word, said clearly, started within three seconds? |
The 20-second check after every task
Answer yes or no. 1. I started within a second of the tone or the microphone opening. 2. I never went back to fix a word. 3. I had no silence near three seconds. 4. I finished before the time ran out. 5. For open tasks: I gave the main point, a relationship and a conclusion. 6. For open tasks: every sentence was about this task, not a stock phrase. Five or six yeses: move on. Four or fewer: redo the task tomorrow before opening a new one.
Read Aloud: 8 passages
Each passage is 50 to 60 words, inside the 60-word limit Pearson describes. Start a 35-second timer and prepare with the routine from the Read Aloud guide: read for meaning, find the hard words, mark the chunks, rehearse the first chunk. Then read once, recorded, without restarting. The models show the chunk marks (one slash for a short pause, two at a full stop) and the stressed words in capitals.
RA01: Urban heat islands
Read the text aloud.
Cities are often several degrees warmer than the countryside around them, especially at night. Concrete and asphalt absorb heat during the day and release it slowly after sunset. Planners are now testing simple remedies, such as lighter roofing materials, street trees and small parks, which can lower local temperatures and reduce the demand for air conditioning.
Timer: Prepare: 30 to 40 seconds while the screen counts down. Speak: after the tone, once; the recording window depends on the length of the text.
Traits scored: Content (each replaced, omitted or inserted word counts as one error; the maximum depends on the length of the text), Pronunciation 0 to 5, Oral Fluency 0 to 5. Per the format page it affects Reading and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Read it like this (/ = short pause, // = longer pause at a full stop, CAPITALS = stressed words):
Cities are often several degrees WARMER / than the COUNTRYSIDE around them, / especially at NIGHT. // CONCRETE and ASPHALT / absorb heat during the DAY / and release it SLOWLY after sunset. // Planners are now testing SIMPLE remedies, / such as LIGHTER roofing materials, / street TREES / and small PARKS, / which can lower local TEMPERATURES / and reduce the demand for AIR conditioning.
Let the pitch rise a little on "lighter roofing materials" and "street trees", because the list is not finished, then fall on "air conditioning".
What the scorer rewards: Content counts every word, so read "the countryside around them", not "the countryside"; a dropped "the" or an added "and" is one error each. Oral Fluency rewards an even pace through the three-item list, which is where most readers stall. Say "asphalt" as AS-falt and keep the final t.
RA02: Forecasting the monsoon
Read the text aloud.
Forecasting the monsoon matters to millions of farmers who plan their sowing around the first rains. Meteorologists combine satellite images, ocean temperatures and computer models to estimate when the rains will arrive and how heavy they will be. Even so, local forecasts remain difficult, because a single storm can change conditions within a few hours.
Timer: Prepare: 30 to 40 seconds while the screen counts down. Speak: after the tone, once; the recording window depends on the length of the text.
Traits scored: Content (each replaced, omitted or inserted word counts as one error; the maximum depends on the length of the text), Pronunciation 0 to 5, Oral Fluency 0 to 5. Per the format page it affects Reading and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Read it like this:
FORECASTING the monsoon / matters to MILLIONS of farmers / who plan their SOWING / around the FIRST rains. // METEOROLOGISTS combine SATELLITE images, / OCEAN temperatures / and COMPUTER models / to estimate WHEN the rains will arrive / and how HEAVY they will be. // Even SO, / LOCAL forecasts remain DIFFICULT, / because a SINGLE storm / can change conditions / within a FEW hours.
"Meteorologists" has six syllables: mee-tee-uh-ROL-uh-jists. Say it alone five times in your preparation time before you start the whole text.
What the scorer rewards: The long word at the start of sentence two is the trap. A restart ("meteoro... meteorologists") is a false start, and the Score Guide's top Oral Fluency band allows none, while its Advanced Pronunciation band allows a few minor distortions. A slightly imperfect word usually costs less than going back. Keep moving.
RA03: Sleep and memory
Read the text aloud.
Research on sleep suggests that the brain does not simply rest at night. During deep sleep, it appears to replay and strengthen the information learned during the day. Students who sleep well after studying often recall more the following morning than those who stay awake revising, which is a strong argument against the all-night study session.
Timer: Prepare: 30 to 40 seconds while the screen counts down. Speak: after the tone, once; the recording window depends on the length of the text.
Traits scored: Content (each replaced, omitted or inserted word counts as one error; the maximum depends on the length of the text), Pronunciation 0 to 5, Oral Fluency 0 to 5. Per the format page it affects Reading and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Read it like this:
Research on SLEEP suggests / that the brain does not SIMPLY rest at night. // During DEEP sleep, / it APPEARS to REPLAY and STRENGTHEN / the information learned during the DAY. // Students who sleep WELL after studying / often recall MORE the following morning / than those who stay AWAKE revising, / which is a STRONG argument / AGAINST the all-night study session.
"Learned" here is one syllable, LERND. Link "stay awake" smoothly, as if it were "stay-yawake".
What the scorer rewards: The last sentence runs 28 words with only one comma. Readers run out of breath and pause in the wrong place, for example straight after "than". The scorer listens for phrasing that follows meaning, so take your one real breath before "which"; the other breaks in that sentence are tiny phrase breaks, not breaths.
RA04: What the printing press changed
Read the text aloud.
The printing press did far more than make books cheaper. By allowing identical copies of a text to circulate widely, it helped standardise spelling, spread scientific ideas and encourage public debate. Historians still disagree, however, about how quickly these changes happened and how much they depended on rising literacy rather than on the technology itself.
Timer: Prepare: 30 to 40 seconds while the screen counts down. Speak: after the tone, once; the recording window depends on the length of the text.
Traits scored: Content (each replaced, omitted or inserted word counts as one error; the maximum depends on the length of the text), Pronunciation 0 to 5, Oral Fluency 0 to 5. Per the format page it affects Reading and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Read it like this:
The PRINTING press / did FAR more / than make books CHEAPER. // By allowing IDENTICAL copies of a text / to circulate WIDELY, / it helped STANDARDISE spelling, / spread SCIENTIFIC ideas / and encourage PUBLIC debate. // Historians still DISAGREE, / HOWEVER, / about how QUICKLY these changes happened / and how MUCH they depended on rising LITERACY / rather than on the TECHNOLOGY itself.
"However" sits between two commas, so it gets a small pause on each side and a slight fall in pitch.
What the scorer rewards: Three words carry most of the risk: "identical" (eye-DEN-ti-kul), "circulate" (SUR-kyuh-late) and "literacy" (LIT-uh-ruh-see). Word stress in the wrong place is named in the Score Guide's Pronunciation descriptors, so say each one alone three times during the 30 to 40 seconds of preparation.
RA05: Microfinance and its limits
Read the text aloud.
Microfinance began with a simple idea: people without collateral could still repay small loans if they borrowed in groups that shared responsibility. Early programmes reported high repayment rates and attracted global attention. Later studies were more cautious, finding that small loans helped some households build businesses but rarely lifted entire communities out of poverty on their own.
Timer: Prepare: 30 to 40 seconds while the screen counts down. Speak: after the tone, once; the recording window depends on the length of the text.
Traits scored: Content (each replaced, omitted or inserted word counts as one error; the maximum depends on the length of the text), Pronunciation 0 to 5, Oral Fluency 0 to 5. Per the format page it affects Reading and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Read it like this:
MICROFINANCE began with a SIMPLE idea: / people WITHOUT collateral / could still REPAY small loans / if they borrowed in GROUPS / that SHARED responsibility. // EARLY programmes reported HIGH repayment rates / and attracted GLOBAL attention. // LATER studies were more CAUTIOUS, / finding that small loans helped SOME households build businesses / but RARELY lifted entire communities out of poverty / on their OWN.
The colon after "idea" is a pause with the pitch held level, as if you are about to explain something.
What the scorer rewards: "Collateral" (kuh-LAT-uh-rul) and "responsibility" (six syllables) are long words where a syllable is easy to add or drop, and the Score Guide's lower Pronunciation bands mention syllables "added or missed". The contrast "SOME households ... RARELY ... entire communities" carries the meaning; stress it and the reading sounds understood, not decoded.
RA06: Coral bleaching
Read the text aloud.
Coral reefs cover a tiny fraction of the ocean floor, yet they support an extraordinary variety of marine life. When water temperatures rise, corals expel the algae that give them both colour and food, a process known as bleaching. If temperatures return to normal quickly, many reefs recover; if not, large areas may die.
Timer: Prepare: 30 to 40 seconds while the screen counts down. Speak: after the tone, once; the recording window depends on the length of the text.
Traits scored: Content (each replaced, omitted or inserted word counts as one error; the maximum depends on the length of the text), Pronunciation 0 to 5, Oral Fluency 0 to 5. Per the format page it affects Reading and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Read it like this:
CORAL reefs cover a TINY fraction of the ocean floor, / YET they support / an EXTRAORDINARY variety of marine life. // When water temperatures RISE, / corals EXPEL the algae / that give them BOTH colour AND food, / a process known as BLEACHING. // If temperatures return to normal QUICKLY, / MANY reefs recover; / if NOT, / LARGE areas may DIE.
"Algae" has two syllables (AL-jee and AL-gee are both standard). The semicolon after "recover" is a full pause; then make the contrast clear: "if NOT".
What the scorer rewards: "Extraordinary" is usually said in five syllables, ik-STRAW-din-uh-ree, not spelt out as "extra ordinary". Rehearse it, then keep "an extraordinary variety of marine life" in one breath. "Yet" is the hinge of sentence one; a small pause before it shows the scorer you read the meaning, which is what Oral Fluency phrasing means.
RA07: Antibiotic resistance
Read the text aloud.
Antibiotics transformed medicine in the twentieth century, but their success has created a new problem. Each time bacteria are exposed to a drug, the few that survive can multiply and pass on their resistance. Doctors therefore ask patients to take antibiotics only as prescribed and never to demand them for viral infections, which they cannot treat.
Timer: Prepare: 30 to 40 seconds while the screen counts down. Speak: after the tone, once; the recording window depends on the length of the text.
Traits scored: Content (each replaced, omitted or inserted word counts as one error; the maximum depends on the length of the text), Pronunciation 0 to 5, Oral Fluency 0 to 5. Per the format page it affects Reading and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Read it like this:
ANTIBIOTICS transformed MEDICINE / in the TWENTIETH century, / but their SUCCESS / has created a NEW problem. // Each time BACTERIA are exposed to a drug, / the FEW that survive / can MULTIPLY / and pass on their RESISTANCE. // Doctors THEREFORE ask patients / to take antibiotics ONLY as prescribed / and NEVER to demand them for VIRAL infections, / which they CANNOT treat.
"Twentieth" ends in a th sound; do not turn it into "twentiet". "Therefore" starts with the voiced th of "the", not a d.
What the scorer rewards: This passage tests the th sound twice ("twentieth", "therefore") and the cluster in "transformed". Pronunciation is judged against what most regular speakers of English produce, so a th that becomes t or d is the kind of consistent sound error the Score Guide's Good and Intermediate bands describe. Practise both words until they are steady.
RA08: Remote work after the office closed
Read the text aloud.
When offices closed suddenly, many organisations discovered that a large share of their work could be done from home. Productivity did not collapse, as some managers had feared. However, new employees often found it harder to learn informal skills, such as how decisions are really made, which is why many firms now prefer a hybrid arrangement.
Timer: Prepare: 30 to 40 seconds while the screen counts down. Speak: after the tone, once; the recording window depends on the length of the text.
Traits scored: Content (each replaced, omitted or inserted word counts as one error; the maximum depends on the length of the text), Pronunciation 0 to 5, Oral Fluency 0 to 5. Per the format page it affects Reading and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Read it like this:
When offices closed SUDDENLY, / MANY organisations discovered / that a LARGE share of their work / could be done from HOME. // Productivity did NOT collapse, / as SOME managers had FEARED. // HOWEVER, / NEW employees often found it HARDER / to learn INFORMAL skills, / such as how decisions are REALLY made, / which is WHY many firms now prefer / a HYBRID arrangement.
"Organisations" is or-guh-ny-ZAY-shunz, five syllables. "Hybrid" is HY-brid.
What the scorer rewards: The negative "did NOT collapse" carries the meaning of sentence two. Readers who stress "collapse" instead sound as if they mean the opposite. Sentence-level stress is part of the Score Guide's top two Pronunciation bands, so rehearse the stress pattern as well as the words.
Repeat Sentence: 8 sentences
You need someone, or something, to read each sentence to you once: a study partner at normal speed, or any text-to-speech reader. The sentence is hidden behind a click so you do not see it first. When it ends, start speaking at once and give yourself 15 seconds. Check the recording against the sentence and decide which Content band you reached: 3 for every word in order, 2 for at least half, 1 for less. The test has 10 to 12 of these, so add the ten sentences on the Repeat Sentence page for a full set.
RS01: The library closing early (12 words)
Listen once, then repeat the sentence exactly.
Show the sentence (for the person reading it to you)
Audio: "The library will close early on Friday because of the staff meeting."
Timer: Listen: 3 to 9 seconds of audio, played once. Speak: 15 seconds, starting as soon as the audio ends.
Traits scored: Content 0 to 3 (3 = all words in the right order; 2 = at least half; 1 = less than half; 0 = almost nothing), Pronunciation 0 to 5, Oral Fluency 0 to 5. Feeds Listening and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Say exactly: "The library will close early on Friday because of the staff meeting."
Hear it as three meaning chunks: the library will close early / on Friday / because of the staff meeting. Picture a locked library door with a Friday calendar page taped to it. The picture holds the first two chunks, and the reason follows.
If the middle slips, say "The library will close early ... because of the staff meeting" without stopping. Most words stay in their original order, which is what partial credit counts.
What the scorer rewards: Content counts replacements, omissions and insertions only. "Due to" for "because of" replaces two words. Hesitations are ignored for Content, per the Score Guide, but they still lower Oral Fluency. Copy the speaker's falling pitch on "meeting".
RS02: Uploading an assignment (12 words)
Listen once, then repeat the sentence exactly.
Show the sentence (for the person reading it to you)
Audio: "Please upload your assignment to the online portal before midnight on Thursday."
Timer: Listen: 3 to 9 seconds of audio, played once. Speak: 15 seconds, starting as soon as the audio ends.
Traits scored: Content 0 to 3 (3 = all words in the right order; 2 = at least half; 1 = less than half; 0 = almost nothing), Pronunciation 0 to 5, Oral Fluency 0 to 5. Feeds Listening and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Say exactly: "Please upload your assignment to the online portal before midnight on Thursday."
Chunks: please upload your assignment / to the online portal / before midnight on Thursday. It is an instruction, so it starts high on "please" and falls on "Thursday".
Memory hook: what, where, when. The assignment is the what, the portal is the where, midnight on Thursday is the when. If you hold the three content words, the small words usually follow.
The th in "Thursday" is made with the tongue between the teeth, not a t.
What the scorer rewards: Small words are where Content is lost: "to the", "on". Many candidates say "in the online portal" or "before Thursday midnight", which keeps the meaning but changes the words and their order. Partial credit counts the exact sequence, so practise the small words, not only the nouns.
RS03: Lecture slides online (12 words)
Listen once, then repeat the sentence exactly.
Show the sentence (for the person reading it to you)
Audio: "Students who missed the lecture can download the slides from the website."
Timer: Listen: 3 to 9 seconds of audio, played once. Speak: 15 seconds, starting as soon as the audio ends.
Traits scored: Content 0 to 3 (3 = all words in the right order; 2 = at least half; 1 = less than half; 0 = almost nothing), Pronunciation 0 to 5, Oral Fluency 0 to 5. Feeds Listening and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Say exactly: "Students who missed the lecture can download the slides from the website."
Chunks: students who missed the lecture / can download the slides / from the website.
The first chunk is a subject with a relative clause; hold it as one picture, the students who were absent. Then the action, then the place.
The -ed in "missed" sounds like a t: MIST. Keep it. "Students who miss the lecture" changes a word.
What the scorer rewards: Past-tense endings and final consonant clusters are an easy loss for speakers of languages whose words usually end in a vowel, such as Telugu, Tamil and Malayalam. "Missed" and "slides" both end in clusters. Say the sentence once slowly with every ending, then at normal speed.
RS04: Findings at a conference (10 words)
Listen once, then repeat the sentence exactly.
Show the sentence (for the person reading it to you)
Audio: "The findings were presented at an international conference last spring."
Timer: Listen: 3 to 9 seconds of audio, played once. Speak: 15 seconds, starting as soon as the audio ends.
Traits scored: Content 0 to 3 (3 = all words in the right order; 2 = at least half; 1 = less than half; 0 = almost nothing), Pronunciation 0 to 5, Oral Fluency 0 to 5. Feeds Listening and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Say exactly: "The findings were presented at an international conference last spring."
Chunks: the findings were presented / at an international conference / last spring.
Stress the content words: FINDINGS, preSENTed, interNATional CONference, LAST SPRING. The verb "presented" is pri-ZEN-tid; the noun "present" is PREZ-unt. Getting word stress right is one of the details the Pronunciation descriptors list.
At ten words this is one of the shorter items, so aim for the full Content score of 3.
What the scorer rewards: In natural speech "at an international" links together, close to "a-tuh-ninternational". Copy that linking rather than separating each word: the Score Guide's top Pronunciation band mentions assimilation and deletions "appropriate to continuous speech". Do not add "the" before "last spring"; that is an insertion.
RS05: Returning lab equipment (12 words)
Listen once, then repeat the sentence exactly.
Show the sentence (for the person reading it to you)
Audio: "All laboratory equipment must be returned to the store room after use."
Timer: Listen: 3 to 9 seconds of audio, played once. Speak: 15 seconds, starting as soon as the audio ends.
Traits scored: Content 0 to 3 (3 = all words in the right order; 2 = at least half; 1 = less than half; 0 = almost nothing), Pronunciation 0 to 5, Oral Fluency 0 to 5. Feeds Listening and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Say exactly: "All laboratory equipment must be returned to the store room after use."
Chunks: all laboratory equipment / must be returned / to the store room after use.
"Laboratory" is luh-BOR-uh-tree in British English; LAB-ruh-tor-ee is common in American English. Either is understood. Pick one and say it smoothly.
Memory hook: a rule printed on a lab wall. Rules start with "all" and contain "must".
What the scorer rewards: This is a rule sentence, so "must be" carries the meaning. "Should be" is one replaced word; dropping "be" is one omission. The Score Guide still gives 2 of 3 for Content when at least half the words come out in the right order, so finish the sentence even after a slip.
RS06: Comparing two theories (11 words)
Listen once, then repeat the sentence exactly.
Show the sentence (for the person reading it to you)
Audio: "The professor suggested that we compare both theories in our essays."
Timer: Listen: 3 to 9 seconds of audio, played once. Speak: 15 seconds, starting as soon as the audio ends.
Traits scored: Content 0 to 3 (3 = all words in the right order; 2 = at least half; 1 = less than half; 0 = almost nothing), Pronunciation 0 to 5, Oral Fluency 0 to 5. Feeds Listening and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Say exactly: "The professor suggested that we compare both theories in our essays."
Chunks: the professor suggested / that we compare both theories / in our essays.
"Theories" starts with th: THEER-eez. "Suggested" is suh-JES-tid in British English.
Memory hook: who (the professor), the advice (compare both theories), where (in our essays). The plural "essays" matters; "essay" is a replaced word.
What the scorer rewards: The trap is the plural endings in "theories" and "essays". Under time pressure many candidates drop the final z sound. The Score Guide's Good Pronunciation band mentions consonants "regularly distorted, omitted or mispronounced"; keep every final sound and you stay above it.
RS07: Programme deadline (12 words)
Listen once, then repeat the sentence exactly.
Show the sentence (for the person reading it to you)
Audio: "Applications for the summer research programme close at the end of March."
Timer: Listen: 3 to 9 seconds of audio, played once. Speak: 15 seconds, starting as soon as the audio ends.
Traits scored: Content 0 to 3 (3 = all words in the right order; 2 = at least half; 1 = less than half; 0 = almost nothing), Pronunciation 0 to 5, Oral Fluency 0 to 5. Feeds Listening and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Say exactly: "Applications for the summer research programme close at the end of March."
Chunks: applications for the summer research programme / close / at the end of March.
"Close" here is a verb with a z sound: KLOHZ. It is the shortest chunk and the easiest to lose, because it is buried between two long phrases.
Memory hook: a deadline poster. What (applications), for what (the summer research programme), deadline (the end of March).
What the scorer rewards: Long noun phrases ("summer research programme") take most of your memory and the verb disappears. Content is counted in sequence, so "Applications for the summer programme close at the end of March" still keeps most words. Never stop to search for the missing word: three seconds of silence ends the recording.
RS08: How group projects are assessed (10 words)
Listen once, then repeat the sentence exactly.
Show the sentence (for the person reading it to you)
Audio: "Group projects will be assessed on both content and presentation."
Timer: Listen: 3 to 9 seconds of audio, played once. Speak: 15 seconds, starting as soon as the audio ends.
Traits scored: Content 0 to 3 (3 = all words in the right order; 2 = at least half; 1 = less than half; 0 = almost nothing), Pronunciation 0 to 5, Oral Fluency 0 to 5. Feeds Listening and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Say exactly: "Group projects will be assessed on both content and presentation."
Chunks: group projects / will be assessed / on both content and presentation.
Stress: GROUP PROjects, will be aSSESSED, on BOTH CONtent and presenTAtion. "Content" as a noun is CON-tent, stress on the first syllable.
This is a ten-word item. Use it to practise starting the instant the microphone opens, with no "er" before the first word.
What the scorer rewards: A late start or a filler does not cost Content (the Score Guide says leading material is ignored for Content), but it does count against Oral Fluency. A clean start and an even pace through "content and presentation" is what lifts Oral Fluency from 3 towards 4 or 5.
Describe Image: 7 images
The images are described in words for now; drawn versions are being made. Read the description, sketch it on paper in 30 seconds if that helps, then start the real clock: 25 seconds to plan with the four-move frame (topic, main feature, two relationships, implication), 40 seconds to speak. The Describe Image guide has the frame for each chart type and ten more images. DI06 is a map, the kind of image PTE Core's format page uses as its example.
DI01: Bar chart: study and paid work by year of degree
Describe the image in detail.
Image: A bar chart titled "Average weekly hours of independent study and paid work, by year of degree", with two bars for each year. Study: Year 1, 10 hours; Year 2, 12; Year 3, 14; Year 4, 20. Paid work: Year 1, 6 hours; Year 2, 9; Year 3, 11; Year 4, 5. (Data invented by TalkDrill for practice.)
Timer: Prepare: 25 seconds. Speak: 40 seconds, after the tone.
Traits scored: Content 0 to 6 in PTE Academic (scored by AI, then reviewed by a human expert before the task score is final) or 0 to 5 in PTE Core, then Pronunciation 0 to 5 and Oral Fluency 0 to 5 (called Fluency in Core). Feeds Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
This bar chart compares the average number of hours students spend each week on independent study and on paid work in each year of a four-year degree. The clearest pattern is that study time rises every year, from ten hours in the first year to twenty in the final year. Paid work also grows at first, from six hours to eleven by the third year, but then falls sharply to only five hours in the final year. This suggests that final-year students cut back on part-time jobs to concentrate on their studies, probably because of dissertations and final examinations.
What the scorer rewards: The Score Guide's top Content band asks for the image described "fully and accurately" with the relationships between features expanded. Here the relationship is the crossover: paid work falls just as study peaks. Eight numbers read as a list is what its band 1 describes: points "without description or elaboration".
DI02: Line graph: museum visitors over ten years
Describe the image in detail.
Image: A line graph titled "Visitors to the City Science Museum, 2016 to 2025 (thousands)". 2016: 120. 2018: 150. 2020: 40. 2021: 70. 2023: 160. 2025: 210. (Data invented by TalkDrill for practice.)
Timer: Prepare: 25 seconds. Speak: 40 seconds, after the tone.
Traits scored: Content 0 to 6 in PTE Academic (scored by AI, then reviewed by a human expert before the task score is final) or 0 to 5 in PTE Core, then Pronunciation 0 to 5 and Oral Fluency 0 to 5 (called Fluency in Core). Feeds Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
The line graph shows the number of visitors to the City Science Museum between 2016 and 2025, in thousands. Overall, visitor numbers grew over the decade, but not steadily. They rose from 120,000 in 2016 to 150,000 in 2018, then dropped dramatically to just 40,000 in 2020, which may reflect a period when the museum was closed or entry was restricted. After that, the figures recovered quickly, passing the earlier peak with 160,000 in 2023 and reaching a high of 210,000 in 2025. If this trend continues, the museum may need more space or longer opening hours.
What the scorer rewards: On a line graph, the shape matters more than every point. Give the overall trend first, then the one dramatic change, then the end point. Pearson's format page says the best responses include "possible developments, conclusions or implications", so a closing sentence about what may happen next earns its five seconds.
DI03: Pie chart: how students travel to campus
Describe the image in detail.
Image: A pie chart titled "How students at Riverside College travel to campus". Bus 38 per cent; walking 22 per cent; bicycle 15 per cent; car 14 per cent; metro 11 per cent. (Data invented by TalkDrill for practice.)
Timer: Prepare: 25 seconds. Speak: 40 seconds, after the tone.
Traits scored: Content 0 to 6 in PTE Academic (scored by AI, then reviewed by a human expert before the task score is final) or 0 to 5 in PTE Core, then Pronunciation 0 to 5 and Oral Fluency 0 to 5 (called Fluency in Core). Feeds Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
This pie chart shows how students at Riverside College travel to campus. The bus is by far the most popular option, used by thirty-eight per cent of students, which is more than one in three. Walking comes second at twenty-two per cent, followed by cycling at fifteen per cent. Cars and the metro are the least common, at fourteen and eleven per cent respectively. Taken together, eighty-six per cent of students reach campus without a car, which suggests that the college is well connected and that few students need a parking space.
What the scorer rewards: Pie charts invite a list of five slices. The higher Content bands want relationships: "by far", "more than one in three", "taken together". One combined figure, here 86 per cent without a car, shows you understood the whole image, which is what the Score Guide means by a listener building "a complete mental picture".
DI04: Table: rent in four cities, 2020 and 2025
Describe the image in detail.
Image: A table titled "Average monthly rent for a one-bedroom flat (US dollars)", with four cities and two columns, 2020 and 2025. Northport: 900 and 1,250. Eastvale: 700 and 760. Southbridge: 1,100 and 1,150. Westfield: 500 and 800. (Data invented by TalkDrill for practice.)
Timer: Prepare: 25 seconds. Speak: 40 seconds, after the tone.
Traits scored: Content 0 to 6 in PTE Academic (scored by AI, then reviewed by a human expert before the task score is final) or 0 to 5 in PTE Core, then Pronunciation 0 to 5 and Oral Fluency 0 to 5 (called Fluency in Core). Feeds Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
The table compares the average monthly rent for a one-bedroom flat in four cities in 2020 and 2025, in US dollars. Rents went up everywhere, but at very different speeds. The sharpest increases were in Northport, where rent rose from 900 to 1,250 dollars, and in Westfield, where it climbed from 500 to 800, an increase of sixty per cent. By contrast, Eastvale and Southbridge changed very little, rising by only 60 and 50 dollars. Southbridge was the most expensive city in 2020, but by 2025 Northport had overtaken it, which suggests strong demand for housing there.
What the scorer rewards: Tables tempt you to read every cell aloud, which uses all 40 seconds and leaves no time for comparison. Pick the extremes and one change in ranking. Practise saying numbers like 1,250 smoothly ("twelve hundred and fifty"), because a stumble on a number is a hesitation, and hesitations count against Oral Fluency.
DI05: Process: rainwater harvesting in a house
Describe the image in detail.
Image: A diagram titled "Rainwater harvesting in a house", with six labelled stages joined by arrows. 1. Rain falls on the roof. 2. Gutters carry the water to a downpipe. 3. A first-flush diverter removes the first, dirtiest water. 4. A mesh filter catches leaves. 5. The water is stored in an underground tank. 6. A pump sends it to the toilets and the garden.
Timer: Prepare: 25 seconds. Speak: 40 seconds, after the tone.
Traits scored: Content 0 to 6 in PTE Academic (scored by AI, then reviewed by a human expert before the task score is final) or 0 to 5 in PTE Core, then Pronunciation 0 to 5 and Oral Fluency 0 to 5 (called Fluency in Core). Feeds Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
This diagram illustrates how a house can collect and reuse rainwater, in six stages. The process begins when rain falls on the roof and is carried by gutters to a downpipe. Before the water is stored, it is cleaned in two steps: first, a diverter removes the first and dirtiest flow from the roof, and then a mesh filter catches leaves. The clean water is kept in an underground tank, and finally a pump sends it to the toilets and the garden. Overall, the system reduces the amount of mains water the household needs.
What the scorer rewards: In a process diagram, the sequence is the content. Sequencing words (begins, before, first, then, finally) are the relationships between features for this kind of image. Use the passive where the image shows no person ("is carried", "is stored"); it sounds natural and keeps the focus on the water.
DI06: Map: where three seabirds nest (Core style)
Describe the image in detail.
Image: A map of Kerava Island, a fictional island, titled "Nesting areas of three seabird species". Terns: a long strip along the whole northern coast. Puffins: two small patches on the cliffs of the west coast. Gulls: one large area around the harbour town in the south-east. The centre of the island is forest, with no nesting areas. (Map invented by TalkDrill for practice. The PTE Core format page uses a distribution map of this kind as its Describe Image example.)
Timer: Prepare: 25 seconds. Speak: 40 seconds, after the tone.
Traits scored: Content 0 to 6 in PTE Academic (scored by AI, then reviewed by a human expert before the task score is final) or 0 to 5 in PTE Core, then Pronunciation 0 to 5 and Oral Fluency 0 to 5 (called Fluency in Core). Feeds Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
This map shows where three species of seabird nest on Kerava Island. The terns occupy the largest area, a continuous strip that runs along the entire northern coast. Puffins, by contrast, nest in just two small patches on the cliffs of the west coast. Gulls are concentrated in one large area in the south-east, around the harbour town, perhaps because they feed on waste from fishing boats. Interestingly, none of the species nests in the forested centre of the island, which suggests that all three depend on the sea and the coastline.
What the scorer rewards: On a map, the content is mostly location language: along the northern coast, on the west coast, in the south-east, around, in the centre. Naming the three birds without saying where each one nests leaves the listener without the mental picture the Content descriptors ask for. The guess about the gulls is marked as a guess ("perhaps"), which is how to add an implication without inventing facts.
DI07: Photograph: a weekend farmers' market
Describe the image in detail.
Image: A photograph of an open-air farmers' market on a weekend morning. In the foreground, a stallholder weighs tomatoes on a hanging scale for a customer holding a cloth bag. In the middle, a long row of stalls under green and white umbrellas sells fruit, vegetables and flowers. In the background, a clock tower shows ten o'clock and the street is crowded with shoppers.
Timer: Prepare: 25 seconds. Speak: 40 seconds, after the tone.
Traits scored: Content 0 to 6 in PTE Academic (scored by AI, then reviewed by a human expert before the task score is final) or 0 to 5 in PTE Core, then Pronunciation 0 to 5 and Oral Fluency 0 to 5 (called Fluency in Core). Feeds Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
This photograph shows an open-air farmers' market, apparently on a weekend morning, since the clock tower in the background reads ten o'clock. In the foreground, a stallholder is weighing tomatoes on a hanging scale for a customer who is holding a cloth bag. Behind them, a long row of stalls under green and white umbrellas sells fresh fruit, vegetables and flowers, and the street beyond is crowded with shoppers. Overall, the picture suggests a popular local market where people buy produce directly from growers, and the cloth bag hints that shoppers here try to avoid plastic.
What the scorer rewards: Pearson's Content criteria are the same for a photograph as for a chart: all parts, organised, with an interpretation. Foreground, middle, background is the simplest order for a scene. Use the present continuous for actions ("is weighing") and tie your interpretation to something visible, such as the cloth bag.
Retell Lecture: 4 lectures
On test day you hear up to 90 seconds of lecture audio, sometimes with an image, and can take notes on the erasable noteboard. Pearson's tips say you get about three seconds to look at the image before the audio starts. Here, have a partner read the lecture summary aloud at lecture pace while you take notes, or read it once yourself and cover it. Then 10 seconds to prepare and 40 seconds to speak. PTE Academic only.
RL01: Green roofs: benefits, costs and a proposal
Retell the lecture in your own words.
Lecture (what you would hear, summarised by TalkDrill): A lecturer in urban planning explains green roofs. A green roof is a layer of soil and plants on top of a waterproof membrane. It keeps buildings cooler in summer, so less energy is used for air conditioning. It soaks up rainwater, so drains overflow less in heavy storms. It gives insects and birds a place to live. The drawbacks: it is heavy, so older buildings may need strengthening, and it costs more to install. The lecturer argues that cities should require green roofs on new buildings, where the extra cost is smallest.
Timer: Listen: up to 90 seconds. Prepare: 10 seconds. Speak: 40 seconds, after the tone.
Traits scored: Content 0 to 6 (scored by AI, then reviewed by a human expert before the task score is final), Pronunciation 0 to 5, Oral Fluency 0 to 5. Feeds Listening and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
The lecturer talked about green roofs, which are layers of soil and plants laid over a waterproof membrane on top of a building. The speaker described three main benefits. First, green roofs keep buildings cooler in summer, which cuts the energy used for air conditioning. Second, they absorb rainwater, so drains are less likely to overflow in heavy storms. Third, they create a habitat for insects and birds. However, the lecturer also mentioned two drawbacks: they are heavy and expensive to install. The conclusion was that cities should require green roofs on new buildings, where the extra cost is lowest.
What the scorer rewards: The Score Guide's top Content band asks for the main ideas paraphrased "in own words", connected in "a logical sequence". In 40 seconds the structure that delivers it is: topic, three benefits, the drawbacks, the speaker's conclusion. The conclusion is the part most retells drop, and it is the clearest sign you followed the argument.
RL02: How vaccines train the immune system
Retell the lecture in your own words.
Lecture (summarised by TalkDrill): A biology lecturer explains how vaccines work. A vaccine shows the immune system a harmless version, or a part, of a germ. The body responds by making antibodies and memory cells. If the real germ arrives later, the memory cells recognise it and the response is faster and stronger, so the person is less likely to become seriously ill. Some vaccines need booster doses because protection fades over time. When many people in a community are immune, the germ spreads less easily, which also protects people who cannot be vaccinated.
Timer: Listen: up to 90 seconds. Prepare: 10 seconds. Speak: 40 seconds, after the tone.
Traits scored: Content 0 to 6 (scored by AI, then reviewed by a human expert before the task score is final), Pronunciation 0 to 5, Oral Fluency 0 to 5. Feeds Listening and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
The lecture explained how vaccines protect the body. According to the speaker, a vaccine introduces the immune system to a harmless version, or just a part, of a germ. In response, the body produces antibodies and memory cells. Later, if the person meets the real germ, those memory cells recognise it immediately, so the immune response is faster and stronger and serious illness is less likely. The lecturer added that some vaccines need boosters because protection weakens over time. Finally, the speaker pointed out that when most people are immune, the disease spreads less, which protects those who cannot be vaccinated.
What the scorer rewards: Notice the reporting phrases: explained, according to, added, pointed out. They tell the listener you are retelling someone else's ideas, and they connect the points. The cause and effect chain (memory cells, so a faster response, so less serious illness) is the kind of relationship the Content descriptors reward.
RL03: The default effect in workplace pensions
Retell the lecture in your own words.
Lecture (summarised by TalkDrill): A lecturer in behavioural economics describes the default effect. When employees have to sign up for a workplace pension themselves, many never do, even though they intend to. When they are enrolled automatically and have to opt out instead, far more stay in the scheme. Two reasons: people put off decisions, and they treat the default as a recommendation. The lecturer warns that defaults can also be misused, for example in subscriptions that are hard to cancel, so the people who design choices have an ethical responsibility.
Timer: Listen: up to 90 seconds. Prepare: 10 seconds. Speak: 40 seconds, after the tone.
Traits scored: Content 0 to 6 (scored by AI, then reviewed by a human expert before the task score is final), Pronunciation 0 to 5, Oral Fluency 0 to 5. Feeds Listening and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
The lecture was about the default effect in behavioural economics, and the speaker used workplace pensions as the main example. When employees must join a pension scheme on their own, many fail to do so, even if they plan to. However, when they are enrolled automatically and have to opt out, far more people remain in the scheme. The lecturer gave two reasons for this: people tend to delay decisions, and they see the default option as advice. Finally, the lecturer warned that defaults can be abused, for instance by subscriptions that are difficult to cancel, so designers have an ethical duty.
What the scorer rewards: This lecture has a contrast (opt in against opt out), two causes and a warning. The Score Guide's band 4 describes a response that may "focus on less important details"; spending 40 seconds on the pension example and skipping the warning lands there. Keep the example to one sentence and save time for the conclusion.
RL04: How the shipping container changed trade
Retell the lecture in your own words.
Lecture (summarised by TalkDrill): A history lecturer explains how the standard shipping container changed world trade. Before containers, goods were loaded by hand in sacks and crates, so ships spent days in port and theft was common. Standard steel boxes could be lifted by crane straight from a ship onto a train or a lorry. Loading became much faster and cheaper, ports moved to large deep-water sites, and many dockworkers lost their jobs. The lecturer concludes that the container is a "boring" invention that shaped globalisation as much as any famous technology.
Timer: Listen: up to 90 seconds. Prepare: 10 seconds. Speak: 40 seconds, after the tone.
Traits scored: Content 0 to 6 (scored by AI, then reviewed by a human expert before the task score is final), Pronunciation 0 to 5, Oral Fluency 0 to 5. Feeds Listening and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
The lecturer discussed how the standard shipping container transformed international trade. Before containers were introduced, cargo was loaded by hand in sacks and crates, which meant ships stayed in port for days and goods were often stolen. Standard steel boxes changed this, because cranes could move them directly from a ship onto a train or a lorry. As a result, loading became far quicker and cheaper, ports relocated to large deep-water sites, and many dockworkers lost their jobs. The lecturer concluded that this apparently boring invention shaped globalisation as much as any famous technology.
What the scorer rewards: History lectures are usually organised as before, change, consequences. Mirror that order and the "logical sequence" in the Content descriptors takes care of itself. "As a result" introduces three consequences in one sentence, which saves time and shows how they connect.
Answer Short Question: 5 questions
Have someone read each hidden question once. There is no tone in this task, so answer the moment they finish, in one word or a few. Pearson's format page says there are no extra marks for additional words and that waiting more than three seconds ends the recording. In PTE Academic this task counts for Listening, not Speaking; it is here because you answer it by speaking, in the middle of Part 1.
ASQ01: Who designs buildings?
Answer the question with one word or a few words.
Show the question (for the person reading it to you)
Question (heard once): "What do we call a person who designs buildings?"
Timer: Listen: 3 to 9 seconds, played once. Speak: 10 seconds; there is no tone, so answer as soon as the status shows Recording.
Trait scored: Vocabulary, 1 (appropriate word choice) or 0. Correct or incorrect only. Counts for Listening in PTE Academic, per its Score Guide; the PTE Core Score Guide lists Listening and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Model answer: "An architect."
That is the whole answer. Pearson's format page says there are no extra marks for additional words, so "An architect" scores the same as "A person who designs buildings is called an architect", and the long version only gives you more chances to stumble.
Say it at once. If you wait more than three seconds, the recording stops and the item is marked Completed.
Near miss: "engineer" is a different job and would not be the appropriate word.
What the scorer rewards: Answer Short Question is scored correct or incorrect on how appropriate your words are, so speed and the right word are all that count. In PTE Academic it affects your Listening score, not Speaking. Say the stressed syllable clearly: AR-ki-tekt.
ASQ02: When does the meeting end?
Answer the question with one word or a few words.
Show the question (for the person reading it to you)
Question (heard once): "If a meeting begins at nine o'clock and lasts ninety minutes, what time does it end?"
Timer: Listen: 3 to 9 seconds, played once. Speak: 10 seconds; there is no tone, so answer as soon as the status shows Recording.
Trait scored: Vocabulary, 1 (appropriate word choice) or 0. Correct or incorrect only. Counts for Listening in PTE Academic, per its Score Guide; the PTE Core Score Guide lists Listening and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Model answer: "Half past ten." ("Ten thirty" is equally fine.)
Do the sum while the question is still playing. Ninety minutes is an hour and a half, and nine plus an hour and a half is half past ten.
Questions with numbers in them reward listening for the numbers first. Write "9" and "90" on your erasable noteboard as you hear them if that helps, then answer within three seconds of the audio ending. Do not explain how you worked it out; there are no marks for the working.
What the scorer rewards: Arithmetic questions test listening, not maths: the numbers come fast and only once. The common slip is mishearing "ninety" as "nineteen". Listen for the stress: NINE-ty (stress first) against nine-TEEN (stress second). Answer in the shortest natural form.
ASQ03: Listening to a heartbeat
Answer the question with one word or a few words.
Show the question (for the person reading it to you)
Question (heard once): "Which instrument does a doctor use to listen to a patient's heartbeat?"
Timer: Listen: 3 to 9 seconds, played once. Speak: 10 seconds; there is no tone, so answer as soon as the status shows Recording.
Trait scored: Vocabulary, 1 (appropriate word choice) or 0. Correct or incorrect only. Counts for Listening in PTE Academic, per its Score Guide; the PTE Core Score Guide lists Listening and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Model answer: "A stethoscope."
Say it STETH-uh-skohp, three syllables, stress on the first. The th at the end of the first syllable is voiceless, as in "think".
If the word does not come instantly, do not fill the silence with "it is the, er". Three seconds of silence ends the recording. A quick, clear single word is the best possible answer here.
Near miss: "thermometer" measures temperature, not heartbeat, and would be marked incorrect.
What the scorer rewards: General-knowledge questions in this task use everyday academic and work vocabulary. The Vocabulary trait is 1 for an appropriate word choice and 0 for an inappropriate one, so a clearly pronounced right word is worth exactly as much as a long sentence containing it.
ASQ04: The opposite of import
Answer the question with one word or a few words.
Show the question (for the person reading it to you)
Question (heard once): "What is the opposite of import?"
Timer: Listen: 3 to 9 seconds, played once. Speak: 10 seconds; there is no tone, so answer as soon as the status shows Recording.
Trait scored: Vocabulary, 1 (appropriate word choice) or 0. Correct or incorrect only. Counts for Listening in PTE Academic, per its Score Guide; the PTE Core Score Guide lists Listening and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Model answer: "Export."
One word. As a noun it is EX-port; as a verb it is ex-PORT. Either stress is understood in a one-word answer, but match the question if you can: the speaker said "import" with stress on the first syllable, so "EX-port" sounds natural.
Opposite questions are among the quickest items in the test. Use them to bank time and calm: answer, wait for the status to change, and move on without second-guessing.
What the scorer rewards: Answer Short Question has no tone before recording, per Pearson's format page, so an easy loss on easy items is a late start. Train the habit of speaking the moment the status shows Recording; a correct word after the recording stops scores nothing.
ASQ05: Sides of a hexagon
Answer the question with one word or a few words.
Show the question (for the person reading it to you)
Question (heard once): "How many sides does a hexagon have?"
Timer: Listen: 3 to 9 seconds, played once. Speak: 10 seconds; there is no tone, so answer as soon as the status shows Recording.
Trait scored: Vocabulary, 1 (appropriate word choice) or 0. Correct or incorrect only. Counts for Listening in PTE Academic, per its Score Guide; the PTE Core Score Guide lists Listening and Speaking.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Model answer: "Six."
Say it clearly, with the final ks sound: SIKS. A dropped final consonant can make "six" sound like "sick", and the scorer judges the word you say, not the one you meant.
If you do not know a word in the question, guess quickly rather than staying silent; silence scores nothing. Here, "hex" signals six, as "pent" signals five in pentagon and "oct" signals eight in octagon.
What the scorer rewards: Very short answers expose final consonants. For speakers whose first language rarely ends words in a consonant cluster, "six" is a good drill word. Answer Short Question counts for Listening, but a clear final sound is still what lets the scorer recognise the right word.
Summarize Group Discussion: 4 discussions
On test day you hear three people discuss a topic for up to three minutes, then get 10 seconds to prepare and two minutes to speak. Pearson's format page says the best responses start with the main topic, then introduce subtopics with supporting details, including which speaker said which piece of information. Practise with two or three friends reading the speaker parts, or read the summary once yourself and cover it. Take notes in a three-column grid, one column per speaker. PTE Academic only.
SGD01: Should every lecture be recorded?
Summarise the discussion in your own words. (Discussion written and summarised by TalkDrill.)
Topic: Should every lecture be recorded and put online?
Speaker 1, Maya (student representative): Recordings help students with part-time jobs and anyone revising. She wants every lecture recorded and posted the same day.
Speaker 2, Dr Grace Okafor (lecturer): Agrees recordings help revision, but has seen attendance fall. Students who only watch online ask fewer questions and miss practical demonstrations. She suggests releasing recordings a week later.
Speaker 3, Ravi (student): Commutes two hours each way and says he would come in more if he could watch some sessions online. He proposes recording lectures but keeping weekly seminars in person and compulsory.
How it ends: They agree to record lectures, keep seminars in person and review attendance after one term.
Timer: Listen: up to 3 minutes. Prepare: 10 seconds. Speak: 2 minutes, after the tone.
Traits scored: Content 0 to 6 (scored by AI, then reviewed by a human expert before the task score is final), Pronunciation 0 to 5, Oral Fluency 0 to 5. Feeds Listening and Speaking. PTE Academic only.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
The discussion was about whether all lectures should be recorded and made available online, and the three speakers approached the question from quite different angles.
Maya, the student representative, was strongly in favour. She argued that recordings support students with part-time jobs and help everyone when they revise, so she wanted every lecture recorded and posted on the same day.
Dr Okafor, a lecturer, partly agreed. She accepted that recordings are useful for revision, but said she had noticed attendance dropping. In her view, students who rely only on recordings ask fewer questions and miss the practical demonstrations that happen in the room. For that reason, she suggested a compromise: releasing each recording a week after the lecture.
Ravi, a student, raised a practical point that linked the two positions. Because he travels two hours each way, he said he would actually come to campus more often if he could watch some sessions online. He proposed recording lectures while keeping the weekly seminars face to face and compulsory, so that students still meet their teachers.
In the end, the group reached an agreement that combined these ideas. Lectures will be recorded, seminars will stay in person, and attendance will be reviewed after one term to check for any negative effect.
Overall, the real disagreement was not about recording itself but about how to protect attendance and interaction.
What the scorer rewards: Pearson's format page says the best responses start with the main topic, then introduce subtopics with supporting details "including which speaker said which piece of information". The Score Guide's top band adds the relationships between points of view. Name each speaker's role, report the position with a reporting verb, then say how the views connect.
SGD02: Dividing the work on a group project
Summarise the discussion in your own words. (Discussion written and summarised by TalkDrill.)
Topic: How should a student group divide the work on a research project?
Speaker 1, Priya: Split the report into four sections by each person's strengths. It saves time and makes each person responsible for one part.
Speaker 2, Tom: Worried that separate sections will not fit together and the writing style will be uneven. He wants weekly meetings to write and edit together.
Speaker 3, Ana: Suggests a middle way. Split the research and the middle sections, but write the introduction and conclusion together in one shared document, and set an internal deadline a week before the real one so there is time to review.
How it ends: They adopt Ana's plan, and Tom agrees to coordinate the shared document and the review week.
Timer: Listen: up to 3 minutes. Prepare: 10 seconds. Speak: 2 minutes, after the tone.
Traits scored: Content 0 to 6 (scored by AI, then reviewed by a human expert before the task score is final), Pronunciation 0 to 5, Oral Fluency 0 to 5. Feeds Listening and Speaking. PTE Academic only.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
The three students were discussing how to divide the work on a group research project, and their main concern was balancing speed with quality.
Priya proposed the most efficient option. She suggested splitting the report into four sections according to each person's strengths, arguing that this would save time and make everyone clearly responsible for one part.
Tom, however, was worried about the result. He pointed out that sections written separately often do not fit together and that the style might be uneven from one part to the next. He therefore preferred weekly meetings where the group would write and edit together.
Ana then offered a compromise that addressed both concerns. She agreed with Priya that the research and the middle sections could be divided, but, like Tom, she wanted the introduction and the conclusion written jointly, in a single shared document. She also suggested an internal deadline one week before the real submission date, so that the group would have time to review the whole report.
In the end, the others accepted Ana's plan, and Tom, who had raised the concern about consistency, agreed to coordinate the shared document and the review week.
So the discussion moved from two opposing views, efficiency against consistency, to a combined approach that keeps the speed of dividing the work while protecting the quality of the final report.
What the scorer rewards: This discussion has a classic shape: proposal, objection, compromise, decision. The Score Guide's band 4 describes a response that "focuses more on individual perspectives than relationships between perspectives". Phrases such as "agreed with Priya that" and "like Tom" are what move you above it, because they connect the speakers instead of listing them.
SGD03: Should the library open 24 hours?
Summarise the discussion in your own words. (Discussion written and summarised by TalkDrill.)
Topic: Should the university library open 24 hours a day?
Speaker 1, Mr Chen (head librarian): Night opening means paying for staff and security, and the library's own entry records show very few students stay after midnight.
Speaker 2, Leila (student): In exam periods the library is full by mid-morning. Students with daytime jobs need late hours, and shared houses are too noisy to study in.
Speaker 3, Daniel (student union officer): Suggests 24-hour opening only during exam weeks, with one floor open, card-only entry and a security guard.
How it ends: They agree to trial Daniel's plan in the next exam period and to count night-time visitors.
Timer: Listen: up to 3 minutes. Prepare: 10 seconds. Speak: 2 minutes, after the tone.
Traits scored: Content 0 to 6 (scored by AI, then reviewed by a human expert before the task score is final), Pronunciation 0 to 5, Oral Fluency 0 to 5. Feeds Listening and Speaking. PTE Academic only.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
The discussion focused on whether the university library should stay open twenty-four hours a day, and it set the cost of the service against students' needs.
Mr Chen, the head librarian, was cautious. He explained that opening at night would mean paying for extra staff and security, and he noted that, according to the library's own entry records, very few students remain in the building after midnight. In his view, the demand does not justify the cost.
Leila, a student, disagreed. She described how, during exam periods, the library is already full by mid-morning. She also argued that students who work during the day need late opening hours, and that many shared houses are too noisy for serious study, so the library is the only quiet option.
Daniel, a student union officer, tried to bring the two sides together. He suggested a limited version of the idea: opening around the clock only during exam weeks, with just one floor available, entry by student card only and a security guard on duty. This would meet Leila's point about exam pressure while keeping Mr Chen's costs down.
The group finally agreed to trial Daniel's plan in the next exam period and to count how many students actually use the library at night.
In short, the speakers moved from a disagreement about cost and need to a small, measurable experiment.
What the scorer rewards: When a speaker cites evidence ("the library's own entry records"), report it as theirs ("according to"). That keeps you accurate and shows the relationship between a claim and its support. Leave out the minor details if time is short; the decision and the reason for the trial must stay in.
SGD04: A compulsory internship for business students?
Summarise the discussion in your own words. (Discussion written and summarised by TalkDrill.)
Topic: Should an internship be compulsory for business students?
Speaker 1, Professor Ruiz: Internships give real workplace experience and help graduates find jobs. He proposes a compulsory eight-week internship for credit.
Speaker 2, Aisha (student): Unpaid internships are unfair to students who need to earn, and international students may face limits on work under their visa conditions.
Speaker 3, Sam (careers adviser): The university could approve only paid placements and offer an on-campus consultancy project with real clients as an alternative.
How it ends: Internships stay optional this year but earn credit, approved placements must be paid, and the consultancy project will be piloted.
Timer: Listen: up to 3 minutes. Prepare: 10 seconds. Speak: 2 minutes, after the tone.
Traits scored: Content 0 to 6 (scored by AI, then reviewed by a human expert before the task score is final), Pronunciation 0 to 5, Oral Fluency 0 to 5. Feeds Listening and Speaking. PTE Academic only.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
The three speakers discussed whether business students should be required to complete an internship, and the debate centred on experience versus fairness.
Professor Ruiz strongly supported the idea. He argued that internships give students genuine workplace experience and make it easier for graduates to find jobs, so he proposed a compulsory eight-week internship that would count towards the degree.
Aisha, a student, raised two objections. First, she pointed out that many internships are unpaid, which is unfair to students who need to earn money while they study. Second, she noted that international students may be restricted in how much they can work under their visa conditions, so a compulsory placement could put them in a difficult position.
Sam, a careers adviser, responded to both concerns with practical solutions. He suggested that the university should approve only paid placements, and that it should offer an alternative: an on-campus consultancy project with real clients, which would give similar experience without the visa or money problems.
In the end, they reached a cautious compromise. Internships will remain optional this year but will earn credit, any approved placement must be paid, and the consultancy project will be piloted.
Overall, everyone agreed that work experience is valuable; the disagreement was about making it compulsory before it is fair for all students.
What the scorer rewards: Aisha gives two separate objections. Signposting them ("First ... Second ...") makes your summary easy to follow, which is part of the Score Guide's top band: ideas "well connected and organized in a logical sequence". The closing sentence that names the real disagreement is the synthesis the top band asks for.
Respond to a Situation: 4 situations
You hear and read a short everyday situation, then say what you would say in it. PTE Academic gives 10 seconds to prepare and PTE Core gives 20; both give 40 seconds to speak. Pearson's Core page advises using the preparation time to decide "what main points you will need to convey, and whether you will need to use a formal or informal register", and to speak for the whole 40 seconds if you can. Practise each situation twice: once with the Academic timer and once with the Core timer.
RTS01: Coffee on a classmate's notes
Situation (you hear it and read it on screen): You borrowed a classmate's lecture notes for the weekend and accidentally spilled coffee on several pages. Your classmate needs them for an exam next week. You meet your classmate on Monday morning. What would you say?
Timer: Prepare: 10 seconds (PTE Academic) or 20 seconds (PTE Core). Speak: 40 seconds, after the tone.
Traits scored: in PTE Academic, Content 0 to 6 (scored by AI, then reviewed by a human expert), Pronunciation 0 to 5 and Oral Fluency 0 to 5. In PTE Core, Appropriacy 0 to 3 (register and politeness), Pronunciation 0 to 5 and Fluency 0 to 5.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Hi Sana, thanks again for lending me your notes. I'm really sorry, but I have to tell you something. On Saturday I knocked my coffee over, and it spilled on three or four pages. Most of the writing is still readable, but two pages are badly stained. I feel terrible, especially with your exam next week. I've already typed up the damaged pages from what I could read, and I can print them for you today. If anything is unclear, I'm happy to check it with someone else's notes. Again, I'm so sorry.
What the scorer rewards: The goal is an apology plus a fix. The Score Guide's top Content band asks for the goal met "with full consideration of the context" (the exam next week) and a response that expands "beyond the prompt language". Register: friendly but serious. In PTE Core the same judgement is the Appropriacy trait.
RTS02: Asked to work on the day of a wedding
Situation (you hear it and read it on screen): Your manager has asked you to work this Saturday to finish an urgent report. You have already promised to attend your cousin's wedding that day, and your family planned it months ago. You go to your manager's office. What would you say?
Timer: Prepare: 10 seconds (PTE Academic) or 20 seconds (PTE Core). Speak: 40 seconds, after the tone.
Traits scored: in PTE Academic, Content 0 to 6 (scored by AI, then reviewed by a human expert), Pronunciation 0 to 5 and Oral Fluency 0 to 5. In PTE Core, Appropriacy 0 to 3 (register and politeness), Pronunciation 0 to 5 and Fluency 0 to 5.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Good morning, do you have a minute? It's about working this Saturday. I understand the report is urgent, and I want to make sure it's finished on time. Unfortunately, Saturday is my cousin's wedding, which my family arranged months ago, so I really can't miss it. Could I suggest an alternative? I'd be happy to stay late on Thursday and Friday, or to work on Sunday if that helps. I could also prepare the data section in advance, so whoever covers Saturday has less to do. Would one of those options work for you?
What the scorer rewards: Saying no to a manager is a register test. The polite structure is: acknowledge, decline with a reason, offer alternatives, ask. PTE Core describes Appropriacy as how appropriate your language is "to the person(s) you are addressing", so "Could I suggest" and "Would one of those options work" matter as much as the content.
RTS03: A new neighbour's barking dog
Situation (you hear it and read it on screen): For a week, your new neighbour's dog has been barking late at night, and you cannot sleep before your early shifts. You have never spoken to this neighbour before. You knock on their door on Sunday afternoon. What would you say?
Timer: Prepare: 10 seconds (PTE Academic) or 20 seconds (PTE Core). Speak: 40 seconds, after the tone.
Traits scored: in PTE Academic, Content 0 to 6 (scored by AI, then reviewed by a human expert), Pronunciation 0 to 5 and Oral Fluency 0 to 5. In PTE Core, Appropriacy 0 to 3 (register and politeness), Pronunciation 0 to 5 and Fluency 0 to 5.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Hi, sorry to bother you on a Sunday. I'm Arjun, from number twelve, next door. I don't think we've met properly, so welcome to the street. I wanted to mention something, and I hope you don't mind. Over the last week, your dog has been barking quite a lot late at night, often after eleven. I start work at six, so I've been struggling to sleep. I'm sure you're not aware of it, as it might happen while you're out or asleep. Is there anything we could try, perhaps keeping the dog inside at night? I'd really appreciate it.
What the scorer rewards: First contact with a stranger plus a complaint: open warmly, state the problem with specifics (when, and its effect on you), assume good faith and ask rather than demand. The Score Guide's band 1 includes a response that meets the goal "with no elaboration or support"; a single polite sentence asking them to stop the barking would land there.
RTS04: The wrong laptop was delivered
Situation (you hear it and read it on screen): You ordered a laptop online for your studies, but the company delivered the wrong model, with less memory than the one you paid for. Your classes start in four days. You phone the company's customer service line. What would you say?
Timer: Prepare: 10 seconds (PTE Academic) or 20 seconds (PTE Core). Speak: 40 seconds, after the tone.
Traits scored: in PTE Academic, Content 0 to 6 (scored by AI, then reviewed by a human expert), Pronunciation 0 to 5 and Oral Fluency 0 to 5. In PTE Core, Appropriacy 0 to 3 (register and politeness), Pronunciation 0 to 5 and Fluency 0 to 5.
Show the model answer (record yours first)
Model, written by TalkDrill (not scored by Pearson)
Good afternoon. I'm calling about an order I received yesterday; the order number is four, five, seven, two, one, nine. I ordered the laptop with sixteen gigabytes of memory, but the model that arrived has only eight, so it isn't what I paid for. I need it for my studies, and my classes begin in four days, so the timing is important. Could you arrange an exchange for the correct model with express delivery? I can send the wrong one back today if you email me a returns label. If an exchange isn't possible in time, I'd like a full refund instead. Thank you.
What the scorer rewards: Formal register on the phone: identify the order, state the fault precisely, explain the urgency, request a specific remedy and give a fallback. The Score Guide's top Content band mentions a "persuasive response"; the deadline and the offer to return the item today are what make this persuasive rather than merely polite.
Two test-order sequences: Academic and Core
Once you have practised each type separately, run the tasks in the order Pearson's format pages list them, without pauses between tasks. This builds the stamina of Part 1, which is the longest part of both tests.
| Sequence | Tasks, in order |
|---|---|
| PTE Academic speaking sequence | A 30-second personal introduction (unscored, as a warm-up); RA01 to RA07; RS01 to RS08 plus four sentences from the Repeat Sentence page; DI01 to DI06; RL01 to RL03; ASQ01 to ASQ05; SGD01 and SGD02; RTS01 and RTS02 with the 10-second timer |
| PTE Core speaking sequence | A 30-second personal introduction (unscored); RA02 to RA08; RS01 to RS08 plus four sentences from the Repeat Sentence page; DI06 (the map), then DI01 to DI03, four images in all; RTS01 to RTS04 with the 20-second timer; ASQ01 to ASQ05 |
For PTE Core, the speaking score is what IRCC converts to a Canadian Language Benchmark. On the canada.ca Express Entry language testing page, checked on 21 October 2026, PTE Core Speaking of 68 to 75 is CLB 7 and 84 to 88 is CLB 9. The PTE Core Speaking page covers the full table and Respond to a Situation in depth.
What the model answers are for, and what they are not
Pearson is direct about memorised answers. The Score Guide says, in its Describe Image scoring flow, that "Responses that contain pre-prepared or memorized material are classified as irrelevant responses", and an irrelevant response scores 0 for Content, which stops the scoring. Pearson's format page adds "You must use your own words" for Describe Image, Retell Lecture and Summarize Group Discussion, and four task types get a human review of Content. So the models for those four task types are study material, not scripts. Read Aloud and Repeat Sentence are different: the words are given, and the models show how to say them.
How to use a model without memorising it:
- Take the moves, not the sentences. Notice how each Describe Image model goes topic, main feature, two relationships, implication. Use that order with your own words on a new image.
- Take five phrases. From each model, pick phrases you would not have used ("by far", "over the same period", "the speaker pointed out") and use each one in a new sentence of your own that day.
- Re-record, do not re-read. After reading a model, close it and answer the same task again. The second recording should be better organised but still yours.
- Swap the details. Change one fact in the task (a different number, a different person in the situation) and answer again, so you practise adapting rather than repeating.
Official practice and your next steps
Use Pearson's own material alongside this bank. All of these were checked on 20 or 21 October 2026:
- PTE Academic Scored Practice Tests: three versions, each a different set of questions, which Pearson says use real past test questions and the same scoring algorithm as the test. They are sold, not free; check the page for the current price.
- PTE Core Scored Practice Test: the Core equivalent, also scored with the test's algorithm, per Pearson.
- PTE Academic Speaking and Writing format and PTE Core Speaking and Writing format: the official task list, timings and test tips, with short videos.
- PTE Academic scoring page: where you download the Score Guide with every trait scale.
Then a simple route. This week: one task type a day from this bank, with its guide open. Next week: one test-order sequence, scored with the sheet, and the 79 plus roadmap to turn the gaps into a daily plan. If your recordings show the same sounds going wrong, read PTE Speaking for Hindi, Punjabi and Telugu speakers or PTE Speaking for Nepali and Filipino candidates. PTE runs on demand and, as the dates block on this page shows, you can book as little as 24 hours ahead; the fees, dates and results page has the booking and results details. Book the official Scored Practice Test about a week before your test date, so there is time to act on what it shows. We plan to refresh this bank every quarter; the date at the top of the page shows the last update.