The task and its timing
An image such as a graph, map, process or photo appears. You have 25 seconds to study it, then 40 seconds to describe it after the tone (Pearson). There are five or six items per test and they feed Speaking only.
The Score Guide uses Describe Image as its worked example. Content is scored first. No response, an irrelevant response, or one built on pre-prepared material gets 0 for content, and there is no further scoring. If content is above 0, pronunciation and oral fluency are scored and the three combine into the item score.
Scales and ranges are from Pearson's PTE Academic Test Taker Score Guide (the post-August-2025 edition, copyright 2025, PDF dated 4 March 2026), read on 4 October 2026. The level descriptions are our own summaries, not Pearson's wording.
Content: 0–6
Since August 2025, content is scored by AI and by a human expert; if they disagree, a second human decides (Pearson). The scale looks at coverage, accuracy, how you link features together, and the range of language you use.
| Score | What it typically means | Common reason for losing it |
|---|---|---|
| 6 | Covers the whole image accurately and interprets how its features relate; varied, precise language; a listener could picture all of it | Listing the data without saying what it shows |
| 5 | Main features accurate and some relationships noted but not developed; good variety of language; only small details missing | Spotting a trend but not comparing it with anything |
| 4 | Some accurate simple description and basic links, but main features may be missed; enough language for the basics, with repetition | Spending the time on one corner of the image |
| 3 | Mostly surface description with minor mistakes; narrow, repetitive language; a listener sees parts, not the whole | Reusing the same few phrases for every feature |
| 2 | Little, shallow description with some errors; simple vocabulary dominates | Running out of things to say after the title |
| 1 | Disconnected items or a list of labels, with no real description | Reading out labels and numbers only |
| 0 | Relevant but too limited for a higher score, or no, irrelevant or memorised content (which stops all scoring) | A generic template that could fit any image |
Know your target? Check where you stand
Take a free AI-scored PTE mock and see your score per skill before you book the real test. Free account, no card.
Take a free scored PTE mock →Pronunciation: 0–5
Pronunciation and fluency are only scored if content is above zero, and they are machine-scored only.
| Score | What it typically means | Common reason for losing it |
|---|---|---|
| 5 Highly proficient | An ordinary listener catches every sound without effort, words run together the way they do in natural speech, and both word stress and sentence stress are right. | Stress slips on numbers such as "thirteen" versus "thirty" |
| 4 Advanced | Sounds are clear, with a few small slips in a vowel, consonant or stress pattern that never make a word hard to understand. | Recurring unclear sounds in data words |
| 3 Good | Most sounds are right, but a recurring slip or two leaves the odd word unclear, and some sounds are dropped or blurred in certain positions. | Swallowed endings on plurals and past tenses |
| 2 Intermediate | Some sounds are wrong every time; roughly two thirds of the speech is understandable and the listener has to tune in to the accent; stress sometimes lands in the wrong place. | Rushing through numbers so stress is lost |
| 1 Intrusive | Many sounds are wrong and the accent gets in the way; a listener could miss about a third of the words, and syllables may be added or dropped. | Heavy first-language patterns across most words |
| 0 Non-English | The speech follows another language's sound system; more than half could be unintelligible and many words have the wrong number of syllables. | Speaking too quietly or indistinctly to be recognised |
Oral Fluency: 0–5
Forty seconds is short, so planning what to say first keeps hesitations down.
| Score | What it typically means | Common reason for losing it |
|---|---|---|
| 5 Highly proficient | Smooth rhythm and natural phrasing, with no hesitations, repeated words, restarts or swallowed sounds. | One hesitation while reading a number off the image |
| 4 Advanced | Natural rhythm and sensible phrasing overall; a single slip (a pause, a repeated word or a restart) is tolerated. | A couple of fillers while deciding what to describe next |
| 3 Good | A reasonable pace that may be uneven; more than one hesitation is possible, but most words come in connected phrases, with no long pauses and no choppy delivery. | Uneven pace, fast on the start and slow on the details |
| 2 Intermediate | Speech is uneven, but a longer answer still contains some fluent stretches of a few words, with only a couple of hesitations and at most one long silence. | A long pause while hunting for a word |
| 1 Limited | Irregular phrasing or word-by-word delivery, with several hesitations, repetitions or restarts; longer answers may contain one or two long pauses. | Several restarts and long pauses |
| 0 Disfluent | Slow and effortful, with most words spoken one at a time and more than one long pause. | Long silences between isolated words |
Retell Lecture: same traits, different input
Retell Lecture is still part of PTE Academic in the current guide; the 2025 update added items but removed none. You hear or watch a lecture of up to 90 seconds, get 10 seconds to prepare, then 40 seconds to retell it (Pearson). There are two or three items per test, feeding Listening and Speaking. Content is 0–6 and double-marked by AI and a human; pronunciation and oral fluency use the same 0–5 scales as above.
On the content scale, the top level paraphrases the main ideas in your own words, adds specific important points and links them clearly. The middle levels catch some main ideas but mix in less important details or loosely connected ideas. The low levels repeat isolated words from the lecture without making sense of them, and 0 is related but too limited to score higher.
Practical advice for Describe Image and Retell Lecture
This is our advice, not Pearson's rules. In the preparation time, find the main point the image or lecture makes, two or three key details, and one conclusion or implication. Pearson's own tips recommend organising the answer this way.
Avoid memorised all-purpose templates. The guide treats answers built from pre-prepared material as irrelevant, which means 0 for content and no further scoring.
Use your own words for comparisons and trends rather than repeating labels. Paraphrasing is what separates the upper content levels.
BandLadder's AI scores Describe Image and Retell Lecture practice on content, pronunciation and fluency, but only Pearson issues official scores.
Frequently asked questions
What is the maximum content score for Describe Image?+
6. Pronunciation and oral fluency are each scored 0–5.
Do templates work for PTE Describe Image?+
Pearson says responses relying on pre-prepared or memorised material are irrelevant and get 0 for content, which stops all scoring for that item.
Is Describe Image marked by a human?+
Content is scored by AI and a human expert, with a second human deciding disagreements. Pronunciation and fluency are machine-only.
Was Retell Lecture removed from PTE Academic?+
No. It is still in the current Score Guide with two or three items per test. It is not part of PTE Core.
How long do I get to speak in Describe Image?+
40 seconds, after 25 seconds to study the image.
Keep reading
Related
- PTE Academic scoring criteria overview
- PTE Repeat Sentence scoring criteria
- Free PTE mock test
- PTE predictions
- All 711 guides
Verified against the linked official sources on 4 October 2026. Rules change: confirm on the official page before you book or apply.
© 2026 BandLadder. Written and checked by the BandLadder team. You may quote or cite this page with credit to BandLadder and a link to it; republishing it in full needs our written permission. Content use policy