CELPIP Speaking Task 3 Template (Describing a Scene)

You get 30 seconds to prepare and 60 seconds to speak. The clock starts on its own and there is no pause button, which is why the first sentence has to be automatic rather than invented on the spot.

CELPIP's 2026 Speaking Pro study pack lists the preparation and speaking time for every task.

Only Tasks 1 and 7 get 90 seconds. Everything else is a minute. If you have been practising Task 3 with a 90-second timer because a study group told you to, you have been rehearsing an answer that runs 50 percent long, and on test day the recording will cut you off mid-sentence.

Sixty seconds is roughly 130 to 150 spoken words at a comfortable pace. That is four or five sentences of real content, not ten.

CELPIP describes Task 3 as "Look at an illustration and describe what is happening." The screen itself, as shown in CELPIP's 2026 study pack, reads:

Describe some things that are happening in the picture below as well as you can. The person with whom you are speaking cannot see the picture.

Two phrases in there do a lot of work. "Some things" is permission to be selective. "Cannot see the picture" is the reason your spatial language matters at all: if you say "he is over there", your listener has nothing to attach that to.

Both lines come from CELPIP's own 2026 study pack. Nobody is going to ask you to describe every object in the frame.

CELPIP's 2026 study pack rates two answers. The first picture is a busy cinema lobby, and that answer got Level 8. It names the place and some people (posters, patrons, people buying tickets), but CELPIP's notes say it "doesn't develop ideas very much" and that the speaker struggles to describe specific parts of the scene.

The second picture is a supermarket, and that answer got Level 9. It sets the scene first ("not very busy, about nine people"), then picks a few things a listener can see: only one register is open, lemons are on sale, and two excited kids are picking up watermelons. CELPIP notes that it is organized well with quite a few details. It was cut off before its last idea, and CELPIP still called the response complete.

So Level 9 did not come from describing more of the picture. It came from one clear opening line and a few precise details. CELPIP's own note on that Level 9 answer says a greater range of more precise words would improve it further, which is what the zone templates below give you.

What CELPIP actually rates on Task 3

CELPIP Speaking is rated on four dimensions: Content/Coherence, Vocabulary, Listenability and Task Fulfillment. Those four names come from CELPIP's own Speaking Performance Standards, and they apply to all eight tasks, Task 3 included.

Three things worth knowing about this list, because most CELPIP prep sites get them wrong:

There is no points formula, and CELPIP publishes no weighting per dimension. A rater listens to the whole response and places it on a scale.

There is no separate Grammar score. Grammar is not one of the four dimensions: CELPIP lists "grammar and sentence structure" and "variety of sentence structure" inside Listenability. That does not make grammar irrelevant, because broken tenses make you harder to follow, and that lands on Listenability first. But nobody is scoring your present continuous in isolation.

"Coherence and Cohesion", "Lexical Resource" and "Fluency and Pronunciation" are IELTS terms. If a template promises to fix your "Lexical Resource" on CELPIP, it was written for a different test.

If you want the fuller picture of how the four dimensions turn into a level, we cover it on the CELPIP Speaking score chart.

What do raters look at under each dimension?

This is CELPIP's own list for each dimension, and what it means on Task 3:

Source: CELPIP Speaking Performance Standards, in CELPIP's Speaking Pro study pack (2026).

What strong Task 3 answers do well

These are the three habits that separate a Task 3 answer that lands from one that wanders. They are not the official criteria, they are what the criteria reward.

No. CELPIP's own advice for this task says "You don't need to describe everything." That single line contradicts most of what you will read about this task online.

CELPIP publishes five key strategies for Describing a Scene. Here they are word for word, and where each one sits in the template on this page:

Source: CELPIP Speaking Pro: Target 9+ Study Pack (2026).

Read that order again. General statement first, then a few details, described well. Nowhere does it ask for coverage.

This matters because the most common way people lose marks on Task 3 is a breathless inventory. Sixty seconds of "and there is a tree, and there is a bench, and there is a dog, and there is a bicycle" is a list, not a description, and a rater hears exactly that. It reads as thin on Content/Coherence and thin on Vocabulary at the same time, because listing forces you into the same four words over and over.

Three or four elements, each with an action and an impression, will always beat twelve elements named and abandoned. If you finish at 55 seconds having covered four things properly, you have done the task.

The one thing you should not do is invent. If you cannot tell whether the man is a teacher or a parent, say "he appears to be a teacher" and move on. CELPIP asks you to describe what is happening, and a small, clearly hedged inference is part of describing. A confident wrong fact is not.

Try this on a real Task 3 question, free

Sign up free and answer the first Speaking Task 3 question on the real clock. You get up to three AI-scored tries, with feedback on what to fix before test day.

Answer a free Speaking Task 3 question

Every answer below runs between 134 and 142 words, which is what 60 seconds sounds like at a comfortable pace. Read one aloud with a timer before you trust that number, because most people speak faster than they think under pressure and finish at 45 seconds with nothing left to say.

Each one follows the same four-zone path as the template above. The settings change. The shape does not.

If you want more phrasing options for the slots above, the Task 3 vocabulary and phrase list gives you alternatives for each one, and the park scene word bank further down this page gives you words for nature, people, animals and objects.

Learn the shape of these lines, not the exact words. A sentence repeated word for word often sounds rehearsed, and it may not fit the picture on your screen.

For example, "This is a busy scene inside a [Place]. Everyone seems to be in a rush." would be wrong for CELPIP's own supermarket picture, because that store is quiet. Keep the move and change the words: "This is a supermarket, and it is not very busy. Only one register is open."

A template is useless until you have run it against pictures you have never seen before. That is the whole difficulty of Task 3: the structure is fixed, the vocabulary is not, and the vocabulary is what falls apart when the picture is unfamiliar.

CELPIP publishes a small number of sample illustrations in its free materials, and they are worth doing first because they are the closest thing to the real style. There are not many of them, and once you have described the same three pictures four times each, you are testing your memory rather than your English.

For practicing more, our Task 3 question bank holds exam-like picture prompts, covering everyday settings: markets, waiting rooms, parks, kitchens, building sites, classrooms, transport. Each one runs on the real 30-second and 60-second clocks, and your recording comes back with AI scoring on the four dimensions above, which is faster than waiting for a teacher to listen to it.

How to practise so it actually transfers:

If you would rather work through the whole section in order, the CELPIP Speaking practice hub has all eight tasks in one place.

A park is one of the best pictures to practise on, because one image gives you people, animals, weather, objects and a lot of space to organise. The words below are grouped the way you would move through the picture. Pick two or three from each group and practise saying them in full sentences, not as a list.

Notice the order: the wide view first, then the foreground, the left, the back and the right, and a closing line. That is the same route as the template above, with the new words dropped into it. To practise, take any park picture and describe it out loud in 60 seconds.

The task before this one, Speaking Task 2, talking about a personal experience, is a narrative task and uses a completely different shape. The task after it, Speaking Task 4, making predictions, uses the same illustration you just described, which is why people often run the two together in practice. All of them sit in our full set of CELPIP Speaking templates.