Duolingo English Test · Guides · November 5, 2026 · 8 min read

DET speaking tasks: photo, Read Then Speak, Interactive Speaking and the Speaking Sample, with model answers

Your Speaking subscore on the Duolingo English Test (DET) comes from four tasks, all recorded at your computer: Speak About the Photo, Read, Then Speak, Interactive Speaking and the Speaking Sample. There is no examiner in the room. Since July 2025, Interactive Speaking, a short conversation with an animated character, has replaced the old Listen, Then Speak task, and Read Aloud was removed. You get one chance at each answer: there is no re-recording. A video of your Speaking Sample is sent to every university that receives your results.

How speaking is scored

Answers are scored by AI on six criteria:

Criterion What it looks at
Content relevance, development, effect on the listener
Discourse coherence clear structure and connections between ideas
Fluency speed, natural pauses, how much you rely on fillers, repetition and restarts
Grammar range and accuracy of structures
Lexis range and precision of vocabulary
Pronunciation intelligibility, sounds, word and sentence stress, intonation

Any standard accent is accepted if it is easy to understand. You do not need to sound American or British; an Indian, Pakistani or Nigerian accent is fine when your words are clear. What costs points is long silence, unclear key words, and answers that are off topic or repetitive.

The four tasks at a glance

Task Prep Speaking time Number Adaptive
Speak About the Photo 20 s up to 90 s 1 no
Read, Then Speak 20 s up to 90 s 1 no
Interactive Speaking none up to 35 s per question 1 set of 6–8 questions yes
Speaking Sample 30 s up to 3 min 1 no

You click “Record Now” to start each answer.

Speak About the Photo

You see a real photo (not AI-generated) for 20 seconds, then describe it for up to 90 seconds. Duolingo’s advice: imagine describing it to someone who cannot see it, use all the time, and if you do not know a word, describe what the thing looks like or does instead of skipping it.

Structure: overview → foreground → background → people’s actions and feelings → a guess about the situation.

Model answer (photo: a family cooking together in a small kitchen; about 150 words)

This photo shows a family cooking together in a small, bright kitchen. In the foreground, a woman in a green apron is cutting vegetables on a wooden board, and next to her a little girl is standing on a stool so she can reach the counter. The girl is holding a spoon and seems to be stirring something in a bowl. Behind them, a man is checking a large pot on the stove, and there’s steam coming out of it. On the shelves in the background, I can see jars, plates and a few plants by the window. Everyone looks relaxed and happy; the girl in particular is smiling. I think they might be preparing a meal for a special occasion, maybe a weekend lunch with relatives, because there seems to be more food than one family would usually need. It looks like a warm, busy moment.

Notice the place phrases (in the foreground, next to her, behind them), present continuous for actions, and might, seems to, I think for guesses.

Read, Then Speak

You read a question about an experience or your opinion, often with bullet points, prepare for 20 seconds and speak for up to 90. The question stays on screen.

Prompt: Talk about a place in your town that you like to visit. You should say: where it is, what you do there, and why you like it.

Model answer (about 160 words)

One place I really like in my town is the old public library near the main market. It’s a two-story building with high windows, and it’s only a ten-minute walk from my house. I usually go there on Saturday mornings. I spend most of the time in the reading room upstairs, where I study for my exams or read English newspapers, and sometimes I join the book discussion group that meets once a month. I like it for a few reasons. First, it’s quiet, which is rare in my neighborhood, so I can focus much better than at home. Second, the librarians are very helpful; one of them recommended the novels that improved my reading most. And finally, it’s simply a calm place in a very busy town. Whenever I feel stressed, a couple of hours there helps me relax and get organized again.

Use the bullets as your order. In 20 seconds, plan one detail for each.

Interactive Speaking

A Duolingo character asks you questions aloud, one at a time, on two topics, three or four questions each, 6–8 in total. You hear each question once; the 35-second timer starts when the character stops speaking. Follow-up questions are chosen based on how well you answered and what you said, so the task adapts. It is the only adaptive speaking task.

A 35-second shape: direct answer (1 sentence) → reason (1–2 sentences) → example or detail (1–2 sentences). That is roughly 60–80 words.

Sample exchange, topic: studying

Q1: What subject are you most interested in studying, and why? A: I’m most interested in computer science. I’ve enjoyed solving problems since school, and programming lets me build things that people actually use. For example, last year I made a small app that helps my mother track her shop’s sales, and seeing her use it every day was really motivating.

Q2: What do you think is the hardest part of learning to program? A: I’d say the hardest part is staying patient when something doesn’t work. A program can fail because of one tiny mistake, and finding it can take hours. What helped me was learning to break problems into smaller steps and test each one, so the errors are easier to find.

Q3: Some people say online courses are as good as university classes. What’s your opinion? A: I think online courses are great for specific skills, but they can’t fully replace university. At university you get feedback from professors and you learn by working with classmates. Online, it’s easy to lose motivation because nobody’s checking your progress.

Each answer responds to the exact question, gives a reason, and adds something concrete. If you do not understand a question, Duolingo’s tip is to say something about the topic rather than stay silent.

Speaking Sample

You read a question about an experience or opinion, take up to 30 seconds to plan, and speak for up to 3 minutes. The question stays on screen. Universities see the video, so look at the camera, sit upright and speak as if to an admissions officer.

Prompt: Some students prefer to study alone, while others prefer to study in groups. Which do you prefer, and why?

Model answer, opening and first point (about 170 words; a full answer continues with a second reason and a conclusion)

Personally, I prefer studying in a small group, although I think the best approach combines both. Let me explain why.

My main reason is that explaining ideas to other people helps me understand them much more deeply. When I prepared for my final school exams, I met twice a week with two friends. We took turns teaching each other one chapter. To explain a chemistry topic clearly, I had to really understand it, not just memorize it, and when my friends asked questions I couldn’t answer, I immediately knew which parts I needed to review. In the end, all three of us got better grades than in our mock exams, and I’m convinced the group sessions were a big part of that.

That said, groups don’t work for everything. For reading long texts or doing practice problems, I need silence and my own pace…

Planning in 30 seconds: position → two reasons → one example for each → a closing line. In three minutes, two well-developed reasons with real examples are better than five thin ones. Speak for as long as you have relevant things to say.

Common mistakes

How to improve

Writing tasks follow the same logic of answer, reason, example; see DET writing tasks. For the full list of question types, see the 2026 DET format.

Practice the speaking tasks in the Nomto room

The Interactive Speaking drill in the Nomto DET room asks 6–8 questions on two topics with a 35-second timer and reviews your answers on content, coherence, fluency, grammar, vocabulary and pronunciation, with corrected sentences and a model answer. The photo, Read, Then Speak and the Speaking Sample have their own drills. Three spoken answers are reviewed free with an account; our ratings are estimates, not Duolingo scores.

Common questions

Can I re-record a speaking answer on the DET? No. You get one attempt at each.

Does my accent lower my score? No, if you are easy to understand. Any standard accent is accepted.

How many questions are in Interactive Speaking? 6 to 8, on two topics, with 35 seconds for each answer.

Do universities see my speaking? They see a video of your Speaking Sample from the test you send.

Duolingo English Test is a trademark of Duolingo, Inc. Nomto is independent and is not affiliated with or endorsed by Duolingo.

Also in Español, Bahasa Indonesia, Português (Brasil)