How to use the 30 tasks
This page is a practice bank for the paired Cambridge English Speaking test: 30 tasks written by TalkDrill in the official format, 10 for B1 Preliminary, 10 for B2 First and 10 for C1 Advanced. Each task has the interlocutor's words and the photo or prompt set, the official timer for that part, a partner side so a friend or the TalkDrill AI can play the second candidate, a model performance hidden until you have tried, and a note naming the criterion the task really tests. Below the tasks, the bank is assembled into six full mock tests at the real length, with a self-score sheet built from the Cambridge band descriptors. Most practice pages give you a script for one person. The Speaking test is taken with another candidate, and one of the criteria, Interactive Communication, can only be shown by responding to someone. So every task here has two sides.
- Run it cold first. Start a phone timer, read the task, speak. Record yourself.
- Your partner reads the partner side. In Parts 2 and 3 it tells them what to say back; in Parts 1 and 4 it tells them how the interlocutor follows up.
- Read the examiner note, then open the model. Compare the shape, not the words. The model is one good performance, not the answer.
- Run it again the next day with roles swapped, and score both recordings with the sheet further down.
Three rules hold for every level. First, keep talking until the interlocutor says "Thank you"; the C1 Advanced handbook warns against "closure techniques such as 'That's all!'". Second, in Part 2 do not speak during your partner's minute; the B2 First handbook says the listening candidate "should not speak during their partner's long turn". Third, in Part 3 you do not have to finish: the B2 First handbook says candidates "are not penalised if they fail to reach a negotiated decision".
The official timers for every part
These are the timings from the cambridgeenglish.org exam-format pages and the Speaking sections of the three handbooks for teachers. Set them on a phone before you start. A minute is longer than it feels: the B2 First handbook tells teachers to time Part 2 practice "so that your students get a feel for how long a minute is".
| Part | B1 Preliminary | B2 First | C1 Advanced |
|---|---|---|---|
| Part 1 interview with the interlocutor | 2 minutes | 2 minutes (3 for three) | 2 minutes (3 for three) |
| Part 2 on your own | 3 minutes in all: describe one colour photograph for about 1 minute each | 4 minutes in all: compare two photographs for 1 minute, then about 30 seconds on your partner's photos | 4 minutes in all (6 for three): talk about two of three pictures for 1 minute, then about 30 seconds on your partner's set |
| Part 3 with your partner | 4 minutes in all: a situation and pictures, about 2 to 3 minutes of discussion | 4 minutes in all (5 for three): 15 seconds to read, a 2-minute discussion, a 1-minute decision | 4 minutes in all (6 for three): about 15 seconds to read, a 2-minute discussion, a 1-minute decision |
| Part 4 discussion | 3 minutes, general conversation | 4 minutes (6 for three) | 5 minutes (8 for three) |
| Total per pair | 10 to 12 minutes (15 to 17 for three) | 14 minutes (20 for three) | 15 minutes (23 for three) |
Two things the timer does not show. At B2 and C1 there are always two examiners: the interlocutor runs the test and gives a Global Achievement mark, the assessor says nothing and marks the analytical criteria. The B2 First, C1 Advanced and B1 Preliminary format pages add that one examiner "could be online, examining remotely". And at B1 the Part 2 and Part 3 tasks are different in kind, not just shorter: you describe one photograph instead of comparing two, and in Part 3 the B1 handbook asks you to "discuss all the visual prompts".
No partner? How to run a paired test with the AI
Most candidates prepare alone, which is the wrong way to prepare for a test marked partly on interaction. If you have nobody to practise with, TalkDrill's role-play practice lets you describe the conversation you want and builds a role-play from it, with an AI counterpart who talks back. You need to sign in to use it. Here is the set-up that works for Cambridge tasks:
- Open role-play practice and choose to describe your own situation. Paste: "Cambridge B2 First Speaking, Part 3. You are Candidate B, my partner. Here is the task:" and then the task text from this page.
- Paste the partner side. It tells the AI what positions to take, for example "thinks the football match is too expensive". That gives you something to disagree with.
- Start a phone timer yourself. The AI does not run Cambridge timings. Two minutes for the discussion, then say "Now we have about a minute to decide" and keep going for one more.
- For Part 2, ask the AI to be the interlocutor. It reads the instructions, you speak for a minute, and it asks you the 30-second question about your partner's photos.
- Listen back and score yourself on the Cambridge criteria with the sheet below. The role-play report is TalkDrill's own feedback, not a Cambridge mark.
For turn-taking under pressure, the group discussion mode runs a timed round with several AI participants who hold their own positions and talk over each other. That is harder than a Cambridge Part 3, which is a polite two-person task, but it trains the habit Part 3 rewards: getting into the conversation and building on what was just said. The mock group discussion guide explains how the round works.
With a human partner
Sit side by side, not face to face across a desk; in the test you both face the examiners. One of you holds the timer. Swap roles each task. If your partner's English is weaker than yours, they read the partner side word for word; that still gives you a real person to react to, which is the point.
The 30 tasks at a glance
Tap a task number to jump to it. Tags give the level and the part, so you can work through one part across levels or one level through all its parts. The B1 tasks cover Parts 2 and 3, where B1 candidates most need a partner; the B2 tasks cover all four parts; the C1 tasks cover Parts 2, 3 and 4, and the C1 mocks below add Part 1 questions.
| Task | Level and part | Title | Timer |
|---|---|---|---|
| B1-01 | B1 Part 2 | A family cooking together (one photograph, about 1 minute) | about 1 minute for your photograph (Part 2 is about 3 minutes for the pair) |
| B1-02 | B1 Part 2 | Students studying in a library (one photograph) | about 1 minute for your photograph (Part 2 is about 3 minutes for the pair) |
| B1-03 | B1 Part 2 | An outdoor market (one photograph) | about 1 minute for your photograph (Part 2 is about 3 minutes for the pair) |
| B1-04 | B1 Part 2 | Waiting at a bus stop in the rain (one photograph) | about 1 minute for your photograph (Part 2 is about 3 minutes for the pair) |
| B1-05 | B1 Part 2 | A birthday party in a garden (one photograph) | about 1 minute for your photograph (Part 2 is about 3 minutes for the pair) |
| B1-06 | B1 Part 3 | A first day in a big city: what should the teenager do? | about 2 to 3 minutes of discussion (the whole part takes about 4 minutes) |
| B1-07 | B1 Part 3 | A present for a grandmother’s seventieth birthday | about 2 to 3 minutes of discussion (the whole part takes about 4 minutes) |
| B1-08 | B1 Part 3 | Raising money for a school trip | about 2 to 3 minutes of discussion (the whole part takes about 4 minutes) |
| B1-09 | B1 Part 3 | Helping a new student make friends | about 2 to 3 minutes of discussion (the whole part takes about 4 minutes) |
| B1-10 | B1 Part 3 | A family weekend away with very little money | about 2 to 3 minutes of discussion (the whole part takes about 4 minutes) |
| B2-01 | B2 Part 1 | Where you live and your free time (interview, 2 minutes) | 2 minutes for both candidates together |
| B2-02 | B2 Part 1 | Studies, work and plans (interview, 2 minutes) | 2 minutes for both candidates together |
| B2-03 | B2 Part 2 | Two photographs of people eating (1 minute, then a 30-second response) | 1 minute for the long turn, then about 30 seconds for the partner |
| B2-04 | B2 Part 2 | Two photographs of people learning (1 minute, then a 30-second response) | 1 minute for the long turn, then about 30 seconds for the partner |
| B2-05 | B2 Part 2 | Two photographs of people travelling (1 minute, then a 30-second response) | 1 minute for the long turn, then about 30 seconds for the partner |
| B2-06 | B2 Part 3 | A year out before university (2-minute discussion, 1-minute decision) | 15 seconds to read, 2 minutes to discuss, 1 minute to decide |
| B2-07 | B2 Part 3 | Learning a skill from a teacher or online (2 plus 1 minutes) | 15 seconds to read, 2 minutes to discuss, 1 minute to decide |
| B2-08 | B2 Part 3 | Making a town better for young people (2 plus 1 minutes) | 15 seconds to read, 2 minutes to discuss, 1 minute to decide |
| B2-09 | B2 Part 4 | Discussion: young people, work and independence (4 minutes) | 4 minutes for both candidates together |
| B2-10 | B2 Part 4 | Discussion: towns, facilities and community (4 minutes) | 4 minutes for both candidates together |
| C1-01 | C1 Part 2 | Three photographs of people in demanding jobs (talk about two) | 1 minute on two of the three pictures, then about 30 seconds for the partner |
| C1-02 | C1 Part 2 | Three photographs of people waiting (talk about two) | 1 minute on two of the three pictures, then about 30 seconds for the partner |
| C1-03 | C1 Part 2 | Three photographs of people taking photographs (talk about two) | 1 minute on two of the three pictures, then about 30 seconds for the partner |
| C1-04 | C1 Part 2 | Three photographs of people helping others (talk about two) | 1 minute on two of the three pictures, then about 30 seconds for the partner |
| C1-05 | C1 Part 3 | Why people choose where to live (2-minute discussion, 1-minute decision) | 15 seconds to read, 2 minutes to discuss, 1 minute to decide |
| C1-06 | C1 Part 3 | Reducing stress among university students (2 plus 1 minutes) | 15 seconds to read, 2 minutes to discuss, 1 minute to decide |
| C1-07 | C1 Part 3 | Why people change careers (2 plus 1 minutes) | 15 seconds to read, 2 minutes to discuss, 1 minute to decide |
| C1-08 | C1 Part 3 | A company reducing its environmental impact (2 plus 1 minutes) | 15 seconds to read, 2 minutes to discuss, 1 minute to decide |
| C1-09 | C1 Part 4 | Discussion: pressure, education and responsibility (5 minutes) | 5 minutes for both candidates together |
| C1-10 | C1 Part 4 | Discussion: companies, individuals and the environment (5 minutes) | 5 minutes for both candidates together |
B1 Preliminary: 10 tasks (Parts 2 and 3)
B1 Part 2 is a description task. The B1 handbook says candidates "should keep their descriptions simple, and should not speculate about the context or talk about any wider issues raised by the photographs", and should describe the people and activities "as fully as possible", as if to someone who cannot see the picture: objects, colours, clothing, time of day, weather. If you stop, the interlocutor uses back-up prompts ("Talk about the people", "Talk about the place", "Talk about other things in the photograph"). B1 Part 3 gives you a situation and five pictures; you make and respond to suggestions, discuss the options and agree. The handbook says you are "assessed on their use of appropriate language and interactive strategies, not on their ideas".
B1-01: A family cooking together (one photograph, about 1 minute)
B1 Part 2. Timer: about 1 minute for your photograph (Part 2 is about 3 minutes for the pair).
Interlocutor: Now I’d like each of you to talk on your own about something. I’m going to give each of you a photograph and I’d like you to talk about it. A, here is your photograph. It shows a family cooking at home. B, you just listen. A, please tell us what you can see in the photograph.
The photograph (described in words): a bright kitchen in the daytime. A woman of about forty is stirring a large pot on the cooker. A boy of about ten is standing on a small stool next to her, cutting a tomato with a knife while she watches. An older man with grey hair is sitting at a wooden table reading a recipe on a tablet. There are vegetables, a bowl of eggs and a loaf of bread on the table. A window behind them shows a garden.
Partner side (for a study partner or the AI to play): Interlocutor back-up prompts, used only if Candidate A stops early: “Talk about the people.” “Talk about the place.” “Talk about other things in the photograph.” After about one minute: “Thank you.” Then Candidate B gets a different photograph (for practice, use task b1-02).
What this task really tests: Tests Discourse Management at B1: the Band 3 descriptor asks for “responses which are extended beyond short phrases” with “basic cohesive devices”. Marks go when a candidate lists objects (“a woman, a boy, a table”) and stops at twenty seconds. The B1 handbook says to describe the photograph as if to someone who cannot see it, and not to “speculate about the context”: move through people, actions, place, then time of day and mood, and keep talking until the interlocutor says thank you.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
In this photograph I can see a family in a kitchen. I think it’s the morning or the afternoon because there is a lot of light coming from the window. On the left there’s a woman, maybe the mother, and she’s cooking something in a big pot. Next to her there’s a boy, he’s about ten years old, and he’s standing on a stool because he isn’t tall enough. He’s cutting a tomato and his mother is watching him carefully, perhaps because the knife is sharp. On the right an older man, maybe the grandfather, is sitting at the table. He’s looking at a tablet, I think he’s reading the recipe. On the table there are some vegetables, some eggs and bread. Behind them, through the window, I can see a garden with trees. Everyone looks relaxed and happy, and the boy is concentrating very hard on his tomato.
B1-02: Students studying in a library (one photograph)
B1 Part 2. Timer: about 1 minute for your photograph (Part 2 is about 3 minutes for the pair).
Interlocutor: B, here is your photograph. It shows students in a library. A, you just listen. B, please tell us what you can see in the photograph.
The photograph: a large modern library with tall shelves. In the foreground three students of about eighteen sit around a round table. One girl is typing on a laptop, a boy is writing in a notebook with headphones on, and a second girl is pointing at a page in a thick book and talking to the boy. There are coffee cups and a phone on the table. In the background a librarian is pushing a trolley of books and two people are standing at a computer screen.
Partner side (for a study partner or the AI to play): Interlocutor back-up prompts if needed: “Talk about the people.” “Talk about the place.” “Talk about other things in the photograph.” After about one minute: “Thank you. (Can I have the booklet, please?)”
What this task really tests: Tests Grammar and Vocabulary at B1: present continuous for actions, there is and there are, prepositions of place. The common loss is describing in the present simple (“she write”) or dropping the verb be. Say “she’s typing”, “he’s wearing”, and use left, middle, right, background to organise the picture.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
This photograph shows a library, a very big and modern one, with a lot of shelves full of books. In the middle there are three young people, I think they are university students, and they’re sitting at a round table. The girl on the left is using a laptop, maybe she’s writing an essay. The boy in the middle is wearing headphones and writing in his notebook, so perhaps he’s listening to music while he studies. The other girl is showing him something in a big book and she’s explaining it to him. On the table there are some cups of coffee and a mobile phone. In the background I can see a librarian with a trolley, she’s putting books back on the shelves, and two people are looking at a computer. It looks very quiet, and all three students are concentrating on their work.
B1-03: An outdoor market (one photograph)
B1 Part 2. Timer: about 1 minute for your photograph (Part 2 is about 3 minutes for the pair).
Interlocutor: A, here is your photograph. It shows people at a market. B, you just listen. A, please tell us what you can see in the photograph.
The photograph: a busy street market on a sunny day. In the foreground a woman with a shopping bag is choosing oranges from a stall while the stallholder, a man in an apron, weighs some apples on a scale. Behind them a stall sells cheese and another sells flowers. A small child in a pushchair is holding a balloon. Some people are eating from paper plates near a food van.
Partner side (for a study partner or the AI to play): Interlocutor back-up prompts if needed: “Talk about the people.” “Talk about the place.” “Talk about other things in the photograph.” After about one minute: “Thank you.”
What this task really tests: Tests Pronunciation and Discourse Management together. Candidates who race through a list sound flat; candidates who pause after each sentence sound natural. Group the picture into three zones (front, behind, right) and link them with “behind this”, “on the right”, “near”.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
In this picture I can see a market in the street. It’s a sunny day and there are a lot of people. At the front there’s a woman who is buying fruit. She’s holding a shopping bag and she’s choosing some oranges. The man behind the stall is wearing an apron and he’s weighing some apples for her. Behind this stall there are other stalls. One sells cheese and another one sells flowers, I can see red and yellow flowers. On the right there’s a little child in a pushchair with a balloon, I think it’s a blue balloon. Some people are standing near a food van and they’re eating something from paper plates, maybe sandwiches or hot food. Everybody seems happy, and the market looks very busy and full of colour.
B1-04: Waiting at a bus stop in the rain (one photograph)
B1 Part 2. Timer: about 1 minute for your photograph (Part 2 is about 3 minutes for the pair).
Interlocutor: B, here is your photograph. It shows people waiting for a bus. A, you just listen. B, please tell us what you can see in the photograph.
The photograph: a city bus stop in heavy rain in the early evening. Four people stand under a small shelter. A man in a suit is looking at his watch; a teenage girl with a school bag is sharing an umbrella with a younger boy; an elderly woman is sitting on the bench holding a small dog. The road is wet and car headlights are reflected on it. A bus is visible far down the street.
Partner side (for a study partner or the AI to play): Interlocutor back-up prompts if needed: “Talk about the people.” “Talk about the place.” “Talk about other things in the photograph.” After about one minute: “Thank you.”
What this task really tests: Tests Grammar and Vocabulary: weather and time vocabulary, “maybe” and “I think” for details you cannot be sure of, and a simple future (“they will not wait much longer”). “The people is waiting” is a common slip for Spanish and Italian speakers (our judgement, from the first-language research); practise “people are”.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
This photograph shows a bus stop in a city. It’s raining a lot and I think it’s the evening because it’s getting dark and the cars have their lights on. There are four people under the shelter. On the left a man in a suit is looking at his watch, so maybe he’s late or the bus is late. In the middle a girl, perhaps fifteen years old, is holding an umbrella and sharing it with a smaller boy, maybe her brother. She has a school bag, so I think they’re going home from school. On the right an old woman is sitting on the bench with a small dog on her knees. The road is very wet. At the end of the street I can see a bus coming, so they will not wait much longer. Nobody looks very happy because of the weather, but the children look okay.
B1-05: A birthday party in a garden (one photograph)
B1 Part 2. Timer: about 1 minute for your photograph (Part 2 is about 3 minutes for the pair).
Interlocutor: A, here is your photograph. It shows a party. B, you just listen. A, please tell us what you can see in the photograph.
The photograph: a garden party in the afternoon. A girl of about seven in a yellow dress is about to blow out candles on a cake while six children around a long table clap and sing. Balloons are tied to the chairs. Two adults stand behind the table, one taking a photo with a phone and one holding a jug of juice. A dog is under the table looking at a fallen piece of cake.
Partner side (for a study partner or the AI to play): Interlocutor back-up prompts if needed: “Talk about the people.” “Talk about the place.” “Talk about other things in the photograph.” After about one minute: “Thank you.”
What this task really tests: The B1 handbook says all criteria are assessed in Part 2. Keeping going with very little prompting is the habit to build: if the interlocutor has to use a back-up prompt (“Talk about the people”), you are giving the assessor less to mark. Prepare a fixed route (people, actions, clothes and colours, place, time and weather) and use it on every photograph; the handbook advises describing, not guessing at the wider story.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
In this photograph I can see a birthday party in a garden. It’s a sunny afternoon. In the middle there’s a little girl, I think she’s six or seven, and she’s wearing a yellow dress. She’s going to blow out the candles on her cake. Around the table there are about six other children and they’re clapping and singing, probably “Happy Birthday”. There are balloons on the chairs, red and white ones. Behind the table there are two adults, maybe the parents. The man is taking a photo with his phone and the woman is holding a jug of juice. Under the table there’s a dog and he’s looking at a piece of cake on the grass, I think he wants to eat it. Everybody looks happy, and it looks like a really nice party.
B1-06: A first day in a big city: what should the teenager do?
B1 Part 3. Timer: about 2 to 3 minutes of discussion (the whole part takes about 4 minutes).
Interlocutor: Now, in this part of the test you’re going to talk about something together for about two minutes. I’m going to describe a situation to you.
A sixteen-year-old is visiting a big city for the first time with her uncle. They have one free afternoon. Here are some things they could do. Talk together about the different things they could do, and say which would be most interesting for her.
The pictures (described in words): a science museum; a boat trip on the river; a shopping street; a football match at a stadium; a large park with a lake.
All right? Now, talk together.
Partner side (for a study partner or the AI to play): Candidate B’s likely lines: opens with “Shall we start with the museum?”; thinks the football match is too expensive; likes the boat trip “because you can see everything”; asks “What do you think about the park?”; agrees to the boat trip at the end but suggests the park afterwards if there is time. After 2 to 3 minutes the interlocutor says “Thank you.”
What this task really tests: Tests Interactive Communication: the B1 descriptor asks the candidate to initiate and respond and keep the interaction going. The loss is two monologues in turn. Ask your partner a question at the end of every second turn and react to their idea before adding your own.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
A: Okay, shall we start with the science museum? I think it’s a good idea because it’s interesting and if it rains they can stay inside. What do you think?
B: Yes, maybe, but museums are sometimes boring for teenagers. I prefer the boat trip because you can see the whole city from the river.
A: That’s true, and it’s only one afternoon, so a boat trip is a good way to see a lot of things quickly. What about the football match?
B: I don’t think so. Tickets are expensive and maybe she doesn’t like football.
A: I agree. And the shopping street? She could buy a souvenir.
B: Yes, but she can go shopping in her own town. What do you think about the park?
A: The park is nice and it’s free, but it’s not special for a first visit. So, do we agree that the boat trip is the most interesting?
B: Yes, the boat trip. And if they have time, the park after.
A: Perfect, let’s say the boat trip.
B1-07: A present for a grandmother’s seventieth birthday
B1 Part 3. Timer: about 2 to 3 minutes of discussion (the whole part takes about 4 minutes).
Interlocutor: Now, in this part of the test you’re going to talk about something together for about two minutes. I’m going to describe a situation to you.
A boy and his sister want to buy a present for their grandmother’s seventieth birthday. She likes being at home and spending time with the family. Here are some presents they could buy. Talk together about the different presents and say which would be best.
The pictures: a digital camera; a cookery book; tickets for a concert; a large plant in a pot; a photo album with family pictures.
All right? Now, talk together.
Partner side (for a study partner or the AI to play): Candidate B’s likely lines: starts with the camera and worries that “she might not know how to use it”; likes the concert tickets; changes mind when A says the grandmother prefers staying at home; proposes the photo album; asks “Do you agree?”; accepts the plant as second choice.
What this task really tests: Tests Grammar and Vocabulary in a discussion: simple modals (could, might, should), comparatives (more personal, better) and reasons with because. Using the situation card to justify a choice (“she likes being at home”) is what lifts this from a list to a discussion.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
A: Let’s start with the camera. It’s a nice present, but I’m not sure, because maybe she doesn’t know how to use a digital camera. What do you think?
B: I think the same. And the concert tickets? She could go with the family.
A: That’s a good idea, but the situation says she likes being at home, so maybe she doesn’t want to go out in the evening.
B: You’re right. So what about the photo album with family pictures?
A: I like that one a lot. She can look at it every day, and the family can make it together. The cookery book is okay too, but maybe she already knows how to cook everything.
B: Yes, and the plant? It’s nice for the house.
A: It’s nice, but a plant is a bit normal for a seventieth birthday. In my opinion the album is more personal.
B: I agree. So we choose the photo album?
A: Yes, the photo album is the best, and maybe a small plant with it.
B1-08: Raising money for a school trip
B1 Part 3. Timer: about 2 to 3 minutes of discussion (the whole part takes about 4 minutes).
Interlocutor: Now, in this part of the test you’re going to talk about something together for about two minutes. I’m going to describe a situation to you.
A class wants to raise money for a school trip. They have one month. Here are some ways they could raise money. Talk together about the different ways and say which would be most successful.
The pictures: selling cakes at school; a sponsored walk; a concert in the school hall; washing cars in the car park; selling second-hand books.
All right? Now, talk together.
Partner side (for a study partner or the AI to play): Candidate B’s likely lines: prefers the concert “because parents will pay for tickets”; thinks washing cars is hard work; asks “Have you ever done a sponsored walk?”; worries the cake sale makes little money; agrees on the concert plus a cake sale on the same evening.
What this task really tests: Tests Discourse Management: giving a reason, a doubt and a question in one turn. Candidates who answer “yes, good idea” and wait lose the chance to show extended language. Each turn should carry an opinion plus a because clause.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
A: There are five ideas here. Shall we begin with the cake sale? It’s easy to organise and everybody likes cakes, but I think you don’t earn a lot of money. What’s your opinion?
B: I agree, maybe only a little money. I think the concert is better because the parents will buy tickets and they pay more.
A: Good point. But a concert needs a lot of practice and we only have one month. Is that enough time?
B: Maybe, if the students already play instruments. What about the sponsored walk?
A: I did one last year. It was fun and we got money from family and neighbours. Washing cars is also possible, but it’s hard work and it depends on the weather.
B: And the second-hand books?
A: Students can bring old books from home, so it costs nothing. But I think people don’t pay much for old books.
B: So which is the most successful? I still think the concert.
A: Okay, let’s choose the concert, and we can sell cakes at the concert too. Do you agree?
B: Yes, perfect.
B1-09: Helping a new student make friends
B1 Part 3. Timer: about 2 to 3 minutes of discussion (the whole part takes about 4 minutes).
Interlocutor: Now, in this part of the test you’re going to talk about something together for about two minutes. I’m going to describe a situation to you.
A student has just moved to your country and doesn’t know anyone at the school. He is quite shy. Here are some things he could do to make friends. Talk together about the different things he could do, and say which would be best for him.
The pictures: joining a sports team; going to a language exchange evening; joining the school band; volunteering at an animal shelter; a cooking club.
All right? Now, talk together.
Partner side (for a study partner or the AI to play): Candidate B’s likely lines: suggests the sports team first; when A says the boy is shy, agrees a big team might be difficult; likes the cooking club “because you talk in small groups”; asks “Which one would you choose?”; agrees on the language exchange or the cooking club.
What this task really tests: Tests Interactive Communication: using what the card says about the person (shy, new) to evaluate options is the mark of a real discussion. The loss is ignoring the card and describing each picture. Refer back to the situation at least twice.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
A: Okay, so he’s shy and he doesn’t know anybody. I think the sports team is good for making friends quickly, but a big team can be difficult for a shy person. What do you think?
B: Yes, that’s true. Maybe something smaller is better. What about the language exchange evening?
A: I like that idea, because other students there also want to practise languages, so everybody is a bit nervous and it’s easier to start talking. The band is also nice if he plays an instrument, but we don’t know that.
B: Right. And the animal shelter?
A: Volunteering is a kind thing to do and you meet people, but maybe the other volunteers are adults, not students. I think the cooking club is good because you work in small groups and you can talk while you cook.
B: I agree, cooking club or language exchange. Which one would you choose?
A: For a new student in a new country, the language exchange, because it helps his language too. Shall we say that?
B: Yes, the language exchange is the best.
B1-10: A family weekend away with very little money
B1 Part 3. Timer: about 2 to 3 minutes of discussion (the whole part takes about 4 minutes).
Interlocutor: Now, in this part of the test you’re going to talk about something together for about two minutes. I’m going to describe a situation to you.
A family with two young children wants to go away for the weekend, but they don’t have much money. Here are some things they could do. Talk together about the different things they could do, and say which would be best for the family.
The pictures: camping in the countryside; staying with relatives in another town; a day at the beach; a cycling trip; a day visiting museums in a city.
All right? Now, talk together.
Partner side (for a study partner or the AI to play): Candidate B’s likely lines: likes camping “because it’s cheap and children love it”; worries about the weather; thinks museums are boring for young children; asks “What about staying with relatives?”; agrees camping is best, with the beach as the plan if it rains.
What this task really tests: Tests Discourse Management and Grammar together: the first conditional (“if it rains, camping is terrible”) and comparatives are the natural B1 structures here. The B1 Band 5 Discourse Management descriptor reads “contributions are relevant despite some repetition”, and Band 3 allows “some repetition”; you climb by not saying “it’s cheap” about every option.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
A: Let’s look at the ideas. They have two young children and not much money, so I think the price is important. Camping is cheap and children usually love sleeping in a tent. What do you think?
B: I agree, but it depends on the weather. If it rains, camping is terrible with small children.
A: That’s a good point. The beach has the same problem, but it’s free and it’s only one day. What about staying with relatives?
B: It’s free too, and the children can play with their cousins. But it isn’t really a weekend away.
A: True. A cycling trip is healthy, but young children can’t cycle very far, and the museums in the city might be boring for them, and the city is expensive.
B: So, camping or the beach?
A: I think camping is the best because it’s cheap, it’s a real adventure for the children, and they can go to the beach near the campsite too. Do you agree?
B: Yes, camping is the best. And if the weather is bad, they can visit the relatives instead.
A: Good plan.
B2 First: 10 tasks (Parts 1 to 4)
B2 First tasks follow the pattern in the handbook. Part 1 questions go to each candidate in turn; you do not talk to each other. Part 2 is two photographs and a question printed above them; you are "expected to point out similarities and differences between the photographs and then move on to deal with the question". Part 3 is a central question with five written prompts, then a decision question; the handbook says you "are not expected to discuss all five prompts in the time available but should continue their discussion until asked to stop". Part 4 questions follow the Part 3 topic and, per the handbook, "ask primarily for an evaluation rather than for information". The phrase pages for Part 2 and Part 3 give the language for each move.
B2-01: Where you live and your free time (interview, 2 minutes)
B2 Part 1. Timer: 2 minutes for both candidates together.
Interlocutor: Good morning. My name is Sarah and this is my colleague Tom. And your names are? Can I have your mark sheets, please? Thank you. Where are you from, A? And you, B?
First, we’d like to know something about you.
To A: What do you like most about the place where you live?
To B: Do you prefer to spend your free time indoors or outdoors? (Why?)
To A: Tell us about a friend you spend a lot of time with.
To B: What do you usually do at the weekend?
Partner side (for a study partner or the AI to play): The interlocutor asks each candidate directly and does not invite the partner to comment. Follow-up prompts if an answer is short: “Why?”, “Why not?”, “Tell us more.” Candidate B’s answers should be two or three sentences each; for practice swap roles so both candidates answer all four questions.
What this task really tests: Part 1 tests Grammar and Vocabulary and Pronunciation on familiar topics. The B2 Band 3 descriptor asks for “a good degree of control of simple grammatical forms” while attempting “some complex grammatical forms”, so one “if I had to choose” or “what I value is” per answer is enough. Answers of one sentence give the assessor nothing to mark; in our judgement, answers of thirty seconds are too long for this part.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
What I like most about living in Valencia is probably the mix of city and sea. I can be at the beach in twenty minutes by bike, but there are also plenty of concerts and places to eat, so I never feel bored. If I had to choose one thing, it would be the climate, because it means I can be outdoors most of the year.
A friend I spend a lot of time with is Marta. We’ve known each other since primary school and we’re studying at the same university now, although different subjects. What I value about her is that she’s honest; if I’m making a bad decision she tells me, which not everyone does. We usually meet two or three times a week, often just for a coffee between classes.
B2-02: Studies, work and plans (interview, 2 minutes)
B2 Part 1. Timer: 2 minutes for both candidates together.
Interlocutor: Now, let’s talk about your studies or your work.
To B: Are you studying or working at the moment? Tell us about it.
To A: What did you enjoy most about school?
To B: Is there a skill you would like to learn in the future? (Why?)
To A: Do you think you will live in the same town in ten years’ time? (Why? / Why not?)
Partner side (for a study partner or the AI to play): Follow-up prompts: “Why?”, “Can you tell us more?” The interlocutor moves on after about two minutes in total, so each answer gets roughly twenty to thirty seconds. For practice, the partner asks the questions and times them.
What this task really tests: Tests Grammar and Vocabulary: tense range across present (studying), past (enjoyed), future (will, would like) and present perfect (have started). “I am agree” and “I study since two years” are typical Italian and Spanish transfers (our judgement, from the first-language research); both cost accuracy marks in exactly this part. Keep “for two years” and “I agree” ready.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
I’m in my second year of a degree in economics, and to be honest I enjoy it more than I expected. The first year was mostly theory, but this year we work on real cases, for example analysing why a local company went bankrupt, which makes it feel useful.
A skill I’d really like to learn is programming. Partly it’s practical, because almost every job I’m interested in mentions it, but it’s also curiosity. I’d like to understand how the apps I use every day actually work rather than just using them. I’ve started a short online course, although I haven’t got very far yet.
In ten years’ time I doubt I’ll still be here. I’d like to work abroad for a few years first, probably somewhere in northern Europe, and then come back when I have a family.
B2-03: Two photographs of people eating (1 minute, then a 30-second response)
B2 Part 2. Timer: 1 minute for the long turn, then about 30 seconds for the partner.
Interlocutor: In this part of the test, I’m going to give each of you two photographs. I’d like you to talk about your photographs on your own for about a minute, and also to answer a question about your partner’s photographs. A, it’s your turn first. Here are your photographs. They show people eating in different places. I’d like you to compare the photographs, and say why the people might have chosen to eat in these places.
Question written above the photographs: Why might the people have chosen to eat in these places?
Photograph 1: an extended family of about ten people around a long table at home on a Sunday. Grandparents at one end, small children at the other, serving dishes being passed. Warm indoor light.
Photograph 2: two friends in their twenties sitting on plastic stools at a night street-food stall. Steam from a wok, strings of lights, a crowd behind them, bowls of noodles.
Partner side (for a study partner or the AI to play): After one minute: “Thank you. B, which meal would you prefer to be at? (Why?)” Candidate B answers for about 30 seconds, for example: prefers the street food because it is relaxed and you discover new dishes, but admits the family meal matters more on special occasions. Then the interlocutor gives B two different photographs (use b2-04).
What this task really tests: Tests Discourse Management: a long turn with clear organisation (compare, then answer the question) and cohesive devices (whereas, on the other hand, as for). The handbook says candidates are not expected to describe each picture in detail; a candidate who describes for fifty seconds and never reaches the question loses here.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
Both photographs show people eating together, but the settings are very different. In the first one there’s a large family, three generations I’d say, having a meal at home, probably a Sunday lunch, because the table is full and nobody looks in a hurry. In the second photo, on the other hand, two young friends are eating noodles at a street stall at night, sitting on plastic stools with a crowd behind them. The first meal looks formal and planned, whereas the second seems spontaneous.
As for why they chose these places, I think the family wanted to be together somewhere private, where the grandparents feel comfortable and the children can make noise, and cooking at home is much cheaper for ten people. The friends, however, probably chose the stall because it’s quick, cheap and a bit of an adventure; you can try food you wouldn’t cook yourself, and it’s more about the atmosphere than the meal itself. In both cases, though, the food is really an excuse to spend time together.
B2-04: Two photographs of people learning (1 minute, then a 30-second response)
B2 Part 2. Timer: 1 minute for the long turn, then about 30 seconds for the partner.
Interlocutor: Now, B, here are your photographs. They show people learning in different situations. I’d like you to compare the photographs, and say what might be difficult about learning in these situations.
Question written above the photographs: What might be difficult about learning in these situations?
Photograph 1: a one-to-one violin lesson in a small room. A girl of about twelve is playing while a teacher stands close, adjusting the position of her left hand. Sheet music on a stand.
Photograph 2: a university lecture hall with about three hundred students. The lecturer is a small figure far away at the front next to a projected slide; most students are typing on laptops, a few are looking at their phones.
Partner side (for a study partner or the AI to play): After one minute: “Thank you. A, which of these ways of learning do you think is more effective? (Why?)” Candidate A answers for about 30 seconds: one-to-one is more effective for a practical skill because mistakes are corrected immediately, though lectures suit subjects where you mainly need information.
What this task really tests: Tests Grammar and Vocabulary: speculative language (might, I imagine, probably) and contrast structures. Watch for false friends such as “assist to a lecture” (Spanish and Portuguese “asistir”) and dropped third-person s (“the teacher correct”); the bigger loss is forgetting that the question asks for difficulty, not description.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
Both pictures show learning, but at opposite ends of the scale. The first shows a one-to-one violin lesson: a young girl is playing and her teacher is standing right next to her, correcting the position of her hand. The second shows a huge lecture hall with hundreds of students; the lecturer is a long way from most of them and the students are working on laptops. So the first is personal and practical, while the second is anonymous and mostly about listening.
What might be difficult? In the violin lesson, I imagine the pressure is the main problem. The teacher notices every single mistake, there’s nowhere to hide, and for a child that can be stressful, even though it’s probably the fastest way to improve. In the lecture, the difficulty is the opposite: it’s very easy to switch off. You can’t ask a question without three hundred people turning round, and some students, as you can see, are already on their phones. So one situation is hard because you get too much attention, and the other because you get almost none.
B2-05: Two photographs of people travelling (1 minute, then a 30-second response)
B2 Part 2. Timer: 1 minute for the long turn, then about 30 seconds for the partner.
Interlocutor: A, here are your photographs. They show people travelling in different ways. I’d like you to compare the photographs, and say why the people might have chosen to travel in these ways.
Question written above the photographs: Why might the people have chosen to travel in these ways?
Photograph 1: a packed commuter train at rush hour. People standing shoulder to shoulder, holding rails, looking at phones, a man in a suit with a laptop bag, nobody talking.
Photograph 2: a couple in their sixties cycling along an empty country road with panniers on their bikes, hills and a village in the distance, clear sky.
Partner side (for a study partner or the AI to play): After one minute: “Thank you. B, how do you prefer to travel? (Why?)” Candidate B answers for about 30 seconds: prefers trains for daily life because you can read, but would choose the bike trip for a holiday.
What this task really tests: Tests Discourse Management and Pronunciation: the sentence stress on contrast words (ONLY realistic way, CHOSEN cycling deliberately) is what makes a comparison audible. In our judgement, a B2 Grade B long turn has one clear comparison sentence, then three or four reasons that answer the printed question with modal verbs.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
Both photographs show people on the move, but the mood couldn’t be more different. In the first, we’ve got a crowded commuter train, probably at eight in the morning. Everyone is squeezed together, staring at their phones, and nobody is speaking. In the second, an older couple are cycling along a quiet country road with bags on their bikes, so they’re clearly on a longer trip, maybe a cycling holiday. One picture is about getting somewhere as fast as possible, while the other is about the journey itself.
Why did they choose these ways of travelling? The commuters probably didn’t choose at all, in a sense. The train is likely to be the only realistic way to get to work in a big city; driving would take longer and parking would cost a fortune. The couple, on the other hand, have chosen cycling deliberately. They might be retired and have time, and I suppose they want exercise, fresh air and the freedom to stop in any village they like. It might also be a question of money, since a bike holiday is cheaper than flying somewhere.
B2-06: A year out before university (2-minute discussion, 1-minute decision)
B2 Part 3. Timer: 15 seconds to read, 2 minutes to discuss, 1 minute to decide.
Interlocutor: Now, I’d like you to talk about something together for about two minutes. Some people think young people should spend a year working or travelling before they start university. Here are some things they think about, and a question for you to discuss. First you have some time to look at the task. (15 seconds)
Written task: in the centre, the question “Is it a good idea to take a year out before university?”; around it five prompts: earning money / learning about other cultures / losing the habit of studying / becoming more independent / deciding what to study.
Now, talk to each other about whether it’s a good idea to take a year out before university.
After two minutes: Thank you. Now you have about a minute to decide which of these is the most important thing to consider.
Partner side (for a study partner or the AI to play): Candidate B’s likely lines: opens with “Shall I start?”; strongly in favour of travelling for cultural reasons; dismisses the risk of losing study habits; asks “What about the money side?”; in the decision minute argues for “deciding what to study” but accepts “becoming more independent” if A gives a clear reason.
What this task really tests: Tests Interactive Communication, the criterion Part 3 exists for. The B2 Band 5 descriptor reads “initiates and responds appropriately, linking contributions to those of other speakers” and “negotiates towards an outcome”. Note the moves: an invitation, a concession (“that’s fair”), a link back (“going back to your point”), a transition (“can we move on to”) and a decision offered as a question.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
A: Shall we start with earning money? I think that’s the practical side. A year of work means you start university with some savings, and you also find out what a real job feels like. What’s your view?
B: I agree it’s useful, but for me the cultural side is stronger. If you travel, you meet people who think differently, and that changes you more than a bank balance does.
A: That’s fair. Although, going back to your point, you can’t travel without money, so the two are connected. What worries me a little is losing the habit of studying. After twelve months without an exam, sitting down to read a textbook must be hard.
B: Honestly, I don’t think that’s a big risk. If you want to study, you adapt in a month.
A: Maybe. Can we move on to independence? I’d say this is the real benefit. Whether you work or travel, you’re making decisions without your parents for the first time.
B: Yes, and that links to the last one, deciding what to study. A lot of people choose a degree at seventeen because their friends did.
A: Exactly. So, for the decision: I’d choose becoming more independent, because it helps whatever you study. Would you go with that, or with deciding what to study?
B: I was going to say deciding what to study, but your reason is broader. Let’s say independence.
B2-07: Learning a skill from a teacher or online (2 plus 1 minutes)
B2 Part 3. Timer: 15 seconds to read, 2 minutes to discuss, 1 minute to decide.
Interlocutor: Now, I’d like you to talk about something together for about two minutes. Some people say it is better to learn a new skill from a teacher than from online videos and courses. Here are some things they think about, and a question for you to discuss. First you have some time to look at the task. (15 seconds)
Written task: central question “Is it better to learn a skill from a teacher than online?”; prompts: cost / being able to ask questions / learning at your own pace / staying motivated / meeting other learners.
Now, talk to each other about whether it’s better to learn a skill from a teacher than online.
After two minutes: Thank you. Now you have about a minute to decide which of these things matters most when you choose how to learn.
Partner side (for a study partner or the AI to play): Candidate B’s likely lines: defends online learning on cost and pace; concedes that motivation is a problem (“I have four unfinished courses”); asks A whether it depends on the skill; in the decision minute proposes “staying motivated” as the thing that matters most.
What this task really tests: Tests Interactive Communication and Discourse Management together: the candidate uses the prompts as a route, not a checklist, links two prompts (“part of the same thing”) and uses a personal example without turning it into a monologue. The handbook says there is no need to cover all five prompts; depth over coverage.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
A: Let me start with cost, since it’s the most obvious difference. Online courses are often free or very cheap, while a good teacher can be expensive. For a student that matters a lot. Do you see it the same way?
B: Completely. And connected to that is pace. Online you can repeat a video ten times without feeling embarrassed.
A: True, but you can’t ask the video a question. When I was learning to drive, half of what I learned came from asking “why did that happen?” Doesn’t that depend on the skill, though?
B: I think it does. For something physical, like swimming, you need a person watching you.
A: Right. What about motivation? Be honest, how many online courses have you finished?
B: Not many. I have about four half-finished ones.
A: Same here. A teacher expects you on Tuesday at six, and that expectation is what keeps you going. And the last point, meeting other learners, is part of the same thing; it’s easier to keep going with a group.
B: Agreed. So which matters most when you choose?
A: I’d say motivation, because the cheapest course is worthless if you stop after week two. Would you accept that, or would you put cost first?
B: No, motivation. If the money is really tight, cost decides, but in general motivation.
B2-08: Making a town better for young people (2 plus 1 minutes)
B2 Part 3. Timer: 15 seconds to read, 2 minutes to discuss, 1 minute to decide.
Interlocutor: Now, I’d like you to talk about something together for about two minutes. A town council has some money to spend on improving the town for young people. Here are some things it could spend the money on, and a question for you to discuss. First you have some time to look at the task. (15 seconds)
Written task: central question “How would these things improve life for young people in the town?”; prompts: a skate park / free wifi in the town centre / a cinema / more buses in the evening / a youth club with music rooms.
Now, talk to each other about how these things would improve life for young people in the town.
After two minutes: Thank you. Now you have about a minute to decide which one would benefit the largest number of young people.
Partner side (for a study partner or the AI to play): Candidate B’s likely lines: enthusiastic about the cinema; sceptical about free wifi (“everyone has data”); raises safety for the evening buses; asks “Do you think a skate park is only for a small group?”; in the decision minute agrees on the buses because they help everyone reach everything else.
What this task really tests: Tests Interactive Communication: notice that the candidate evaluates each prompt against the written question (“would it improve life for young people”) rather than against personal taste, and keeps the decision minute for the decision rather than restarting the discussion. Reaching a decision is not required; negotiating towards one is what the descriptor names.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
A: Shall we take them one at a time? A skate park would be great for the people who skate, but I suspect that’s a fairly small group. The wifi, on the other hand, helps everyone, although most teenagers already have data on their phones. What do you think?
B: I’d love a cinema. In a small town there’s nothing to do on a Friday night.
A: I agree, but a cinema is expensive to run, and tickets cost money, so it might not help young people with less money. The buses interest me more. If there are no buses after eight, you can’t go to the cinema or the youth club anyway.
B: That’s a good point, and buses are also about safety. Parents worry about children walking home in the dark.
A: Exactly. And the youth club with music rooms would give people a free place to meet, so for me the buses and the youth club work together.
B: Do you think the skate park is only for a few people, then?
A: Mostly, yes, although it’s cheap and it lasts for years.
B: Okay. For the decision, which benefits the most people?
A: I’d go for the evening buses, because they make every other option reachable and they help people who live outside the centre too. Can we agree on that?
B: Yes, the buses, with the youth club as the next step.
B2-09: Discussion: young people, work and independence (4 minutes)
B2 Part 4. Timer: 4 minutes for both candidates together.
Interlocutor (after the year-out task, questions used in order as appropriate):
Do you think young people today are more or less independent than their parents were at the same age? (Why?)
Some people say that students should work part-time while they study. Do you agree? (Why? / Why not?)
Is it better to travel when you are young or when you are older? (Why?)
Do you think travelling really changes the way people think?
What can young people learn from a first job that they cannot learn at school?
Partner side (for a study partner or the AI to play): The interlocutor directs the questions to one candidate or both, and may ask “What do you think?” or “Do you agree?” to the partner after an answer. Candidate B’s likely positions: thinks young people are less independent because of phones and parents; strongly in favour of part-time work; thinks travel changes people only if they travel alone.
What this task really tests: Part 4 tests Discourse Management and Interactive Communication: justified opinions, developed beyond the minimum, and genuine reactions to the partner. The B2 Band 5 descriptor rewards linking contributions to those of other speakers, so “I’d agree with you up to a point” earns more than a fresh opinion that ignores what was just said.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
On independence, I’d say it depends what we mean. In practical terms my parents were more independent: my mother left home at eighteen and my father was already working. We stay at home longer, partly for financial reasons. On the other hand, we have access to far more information and we make decisions about our future with less pressure from family. So less independent in one sense, more in another.
(To the partner, after B supports part-time work.) I’d agree with you up to a point. A few hours a week teaches you to manage time and money. But I’ve seen friends doing twenty hours in a restaurant and failing exams, so there’s a limit. What matters is whether the job leaves you time to actually study.
And on whether travel changes you, I think it can, but it isn’t automatic. Two weeks in a hotel with people from your own country changes nothing. Living somewhere, dealing with a landlord or a doctor in another language, that’s what shifts how you think.
B2-10: Discussion: towns, facilities and community (4 minutes)
B2 Part 4. Timer: 4 minutes for both candidates together.
Interlocutor (after the town council task):
Do you think local councils should ask young people what they want before spending money? (Why? / Why not?)
Some people say that small towns are better places to grow up than big cities. What do you think?
How important is it for a town to have places where people can meet for free?
Do you think young people today spend too much time at home?
Would you prefer to live in the town where you grew up when you are older? (Why? / Why not?)
Partner side (for a study partner or the AI to play): Follow-up prompts: “What do you think?”, “Do you agree?”, “Why?” Candidate B’s likely positions: councils should consult but young people rarely answer surveys; prefers cities; thinks free meeting places matter more now that cafes are expensive.
What this task really tests: Tests Grammar and Vocabulary at B2: conditional structures (“if they went… they would get”), concession (“but”, “perhaps”) and precise vocabulary (anonymity, consult). The loss here is the one-sentence answer followed by silence; the assessor wants to hear an opinion, a reason and an example every time.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
Yes, I think councils should ask, and not only through a survey that nobody fills in. In my town they put a suggestion box in the library and got about twelve answers in a year. If they went to the schools and asked a class of sixteen-year-olds directly, they would get a much clearer picture, and the young people would feel the decision was partly theirs, so they would look after whatever was built.
On small towns versus cities, I grew up in a town of thirty thousand people and I’d say it was safer and friendlier, but by fifteen it felt small. Everyone knew everything about you. A city gives you anonymity and choice, but you can also feel lost in it. So perhaps a small town until twelve and a city after that, if families could choose, which of course they mostly can’t.
(To the partner.) You said free places matter more now because cafes are expensive, and I think that’s exactly right. If the only place to sit with friends costs four euros a coffee, some people simply stay at home.
C1 Advanced: 10 tasks (Parts 2 to 4)
At C1 Advanced, Part 2 gives you three pictures and two printed questions; you choose two pictures. The C1 handbook says candidates "will always be asked to speculate about something which relates to the focus of the visuals" and "will never be asked to merely describe the visual". Part 3 has the same 2 plus 1 minute shape as B2, and the handbook advises against rushing: "It is better to deal with several of them in depth than to try to deal with all of them superficially." Part 4 lasts 5 minutes. Note that C1 is marked on five analytical criteria, because Grammatical Resource and Lexical Resource are marked separately, and the C1 Band 5 Global Achievement descriptor covers "a wide range of topics, including unfamiliar and abstract ones". The C1 Advanced Speaking page sets out what separates a 180 from a 200.
C1-01: Three photographs of people in demanding jobs (talk about two)
C1 Part 2. Timer: 1 minute on two of the three pictures, then about 30 seconds for the partner.
Interlocutor: In this part of the test, I’m going to give each of you three pictures. I’d like you to talk about two of them on your own for about a minute, and also to answer a question briefly about your partner’s pictures. A, it’s your turn first. Here are your pictures. They show people doing demanding jobs. I’d like you to compare two of the pictures, and say why the people might have chosen to do these jobs, and how difficult it might be to do them well.
Questions written above the pictures: Why might the people have chosen to do these jobs? How difficult might it be to do them well?
Picture 1: a surgeon in a mask and gown in an operating theatre, hands raised, two colleagues beside her, monitors in the background.
Picture 2: a firefighter in full kit carrying a hose towards a smoking building at night, another firefighter signalling.
Picture 3: an air traffic controller alone at a dark console with several radar screens, headset on, coffee cup beside the keyboard.
Partner side (for a study partner or the AI to play): After one minute: “Thank you. B, which of these jobs do you think requires the longest training? (Why?)” Candidate B answers for about 30 seconds: the surgeon, because the training takes more than a decade and mistakes cannot be reversed, although the controller’s job may need a rarer kind of concentration. Then B receives three different pictures (use c1-02).
What this task really tests: Tests Grammatical Resource and Lexical Resource, which C1 Advanced marks separately: the handbook says candidates “will always be asked to speculate” and “will never be asked to merely describe”. The Band 5 Grammatical Resource descriptor is “maintains control of a wide range of grammatical forms”: “may well have been drawn”, “I doubt anyone drifts into”, “having said that”. Our judgement: a 180 answer speculates with “maybe”; a 200 answer grades its certainty.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
I’ll talk about the surgeon and the air traffic controller. What strikes me first is that both are responsible for other people’s lives, yet they look completely different: the surgeon is surrounded by a team under bright lights, whereas the controller is alone in a dark room with nothing but screens and a headset.
As for why they chose these careers, the surgeon may well have been drawn by the prestige and by the very direct sense of helping someone; you operate, and the patient recovers. The controller’s motivation is harder to guess. It could be a fascination with aviation, or simply that it’s well paid and intellectually demanding, and some people are attracted to pressure. I doubt anyone drifts into that job by accident.
In terms of difficulty, I’d argue the controller’s job is harder to do well day after day, because the concentration never lets up and a lapse of ten seconds can be fatal, while the surgeon at least works in short, intense periods with colleagues who can catch a mistake. Having said that, surgery demands years of training that most of us couldn’t sustain.
C1-02: Three photographs of people waiting (talk about two)
C1 Part 2. Timer: 1 minute on two of the three pictures, then about 30 seconds for the partner.
Interlocutor: Now, B, here are your pictures. They show people waiting in different situations. I’d like you to compare two of the pictures, and say why the people might be waiting, and how they might be feeling.
Questions written above the pictures: Why might the people be waiting? How might they be feeling?
Picture 1: an airport departure lounge at night, the board showing several delayed flights, a family asleep across chairs, a businessman pacing with a phone.
Picture 2: a hospital waiting room, a man in his thirties leaning forward with his elbows on his knees, an older woman beside him holding his arm, a nurse at a door.
Picture 3: a long queue outside a shop before dawn, mostly young people in coats with hot drinks, one holding a sign that reads “first in line”, security staff at the door.
Partner side (for a study partner or the AI to play): After one minute: “Thank you. A, which of these people do you think is finding the wait most difficult? (Why?)” Candidate A answers for about 30 seconds: the man in the hospital, because his wait is about a person rather than a thing and he has no control over the outcome.
What this task really tests: Tests Discourse Management at C1: a clear plan (reasons, then feelings), each picture treated under both questions, and cohesive devices beyond “but” and “also” (by contrast, presumably, whereas, although). Talking about the third picture as well is the quickest way to run out of time before the second question; the task says two.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
I’m going to compare the hospital and the queue outside the shop, because they seem to me to be opposite kinds of waiting. In the hospital, a man is sitting with his head down and a woman, perhaps his mother, is holding his arm. They’re clearly waiting for news, possibly about a relative in surgery or a test result. Outside the shop, by contrast, a crowd of young people has queued since before dawn, and they look cheerful, almost festive, with their coffees and a hand-made sign.
So the reasons could hardly be more different. One wait has been forced on the people in the hospital; they didn’t choose it and they have no influence over the outcome, which is presumably why the man looks so tense and the woman is comforting him. The queue, on the other hand, is entirely voluntary. They might be waiting for a new phone or trainers, or tickets, and part of the pleasure is being there together.
As for how they feel, I’d say the hospital pair are anxious but trying to stay composed, whereas the shoppers are excited and probably a little proud of their endurance, although by the time the doors open some of them may be wondering whether it was worth it.
C1-03: Three photographs of people taking photographs (talk about two)
C1 Part 2. Timer: 1 minute on two of the three pictures, then about 30 seconds for the partner.
Interlocutor: A, here are your pictures. They show people taking photographs in different situations. I’d like you to compare two of the pictures, and say what the people might want to capture, and how important the photographs might be to them.
Questions written above the pictures: What might the people want to capture? How important might the photographs be to them?
Picture 1: a tourist taking a selfie in front of a famous monument, a crowd of other tourists doing the same, selfie sticks everywhere.
Picture 2: a wildlife photographer lying in long grass with a very long lens, camouflaged, a deer barely visible in the distance.
Picture 3: a wedding photographer directing a bride and groom on steps, assistants holding a reflector, guests watching.
Partner side (for a study partner or the AI to play): After one minute: “Thank you. B, which of these photographs do you think will be looked at most often in the future? (Why?)” Candidate B answers for about 30 seconds: the wedding photographs, because they will be framed and shown to children, while the selfie disappears into a phone within a week.
What this task really tests: Tests Pronunciation and Discourse Management: the contrast only works if stress lands on the contrasting words (HERSELF being there; INVISIBLE in his own picture). The final concession (“although, to be fair”) is a C1 move that shows flexibility; without it the turn sounds rehearsed.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
Let me take the tourist and the wildlife photographer. On the surface they’re both simply taking a picture, but the attitudes are almost opposite. The tourist is in a crowd, phone at arm’s length, and the monument is really just a backdrop; what she wants to capture is herself being there. The wildlife photographer, meanwhile, has spent hours lying in wet grass, probably alone, waiting for an animal that may never come closer, and he’s invisible in his own picture.
In terms of what they want, I’d suggest the tourist wants proof, something to post and to remember the trip by, and there’s nothing wrong with that. The photographer is after something much rarer: a single moment of behaviour that most people will never see, and he’s prepared to fail most days to get it.
How important are the photos? For the tourist, I suspect not very, in the long run. She’ll take two hundred that day and look at five. For the photographer, that one frame might be the result of a season’s work and could even be how he earns his living, so it matters enormously. Although, to be fair, in twenty years the tourist’s selfie may be the one that makes her smile.
C1-04: Three photographs of people helping others (talk about two)
C1 Part 2. Timer: 1 minute on two of the three pictures, then about 30 seconds for the partner.
Interlocutor: B, here are your pictures. They show people helping others in different ways. I’d like you to compare two of the pictures, and say what the people might find rewarding about helping in these ways, and what might be challenging about it.
Questions written above the pictures: What might the people find rewarding about helping in these ways? What might be challenging?
Picture 1: volunteers in aprons packing bags at a food bank, shelves of tins, a queue of people visible through a door.
Picture 2: a lifeguard on a high chair scanning a crowded beach, whistle in hand, red flag beside the chair.
Picture 3: a young woman tutoring a boy of about nine at a kitchen table, both looking at a maths exercise book, the boy frowning.
Partner side (for a study partner or the AI to play): After one minute: “Thank you. A, which of these kinds of help do you think makes the biggest difference? (Why?)” Candidate A answers for about 30 seconds: the tutoring, because a child who stops being afraid of maths carries that for life, although the food bank meets a more urgent need.
What this task really tests: Tests Discourse Management: both printed questions answered for both pictures within a minute, with a frame (“almost mirror images”) that holds the turn together. If you reach the second question only at fifty seconds, it gets one rushed sentence; plan twenty seconds per move and start the second question by the half-way point.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
I’ll compare the food bank volunteers and the tutor. Both are giving their time rather than money, but the scale is different: the volunteers are helping dozens of strangers they’ll probably never speak to, while the tutor is working with one child she clearly knows, at what looks like a kitchen table.
What might they find rewarding? For the volunteers, I imagine it’s the sense of doing something concrete about a problem that otherwise feels overwhelming. You can’t fix poverty, but you can make sure forty families eat this week. For the tutor, the reward is slower but more personal. The boy is frowning now, but there’ll be a moment when something clicks, and she’ll have caused that.
The challenges are almost mirror images. At the food bank, the difficulty might be emotional distance in the other direction: you see the queue through the door and you know the need is endless, which could be demoralising. The tutor’s challenge is patience. Explaining the same idea a fourth way without showing frustration is genuinely hard, and if the child senses you’re irritated, you’ve lost him. So one is hard because it feels too big, and the other because it’s so small and so close.
C1-05: Why people choose where to live (2-minute discussion, 1-minute decision)
C1 Part 3. Timer: 15 seconds to read, 2 minutes to discuss, 1 minute to decide.
Interlocutor: Now, I’d like you to talk about something together for about two minutes. Here are some places people choose to live, and a question for you to discuss. First you have some time to look at the task. (15 seconds)
Written task: central question “Why might people choose to live in these places?”; prompts: a flat in the city centre / a small village / close to parents and family / in another country / a new housing estate on the edge of town.
Now, talk to each other about why people might choose to live in these places.
After two minutes: Thank you. Now you have about a minute to decide which of these places would be the most difficult to adapt to.
Partner side (for a study partner or the AI to play): Candidate B’s likely lines: begins with the city flat and the pull of work and nightlife; is sceptical about villages (“romantic until you need a hospital”); has lived abroad and says the first year is harder than people admit; asks “Don’t you think the estate is the least interesting but the most common choice?”; in the decision minute argues for “another country”.
What this task really tests: Tests Interactive Communication at C1: the Band 5 descriptor reads “interacts with ease, linking contributions to those of other speakers” and “widens the scope of the interaction and negotiates towards an outcome”. The moves to copy: a genuine follow-up question (“in what way?”), explicit linking of a prompt to the coming decision, and a decision proposed as a question rather than announced.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
A: Shall I kick off with the city-centre flat? I’d say the obvious reasons are work and time: if you can walk to the office you get an hour of your day back. The price you pay is space and noise. How do you see it?
B: Agreed, and nightlife, especially in your twenties. Villages are the opposite: lovely until you need a hospital.
A: That’s a bit harsh, but there’s something in it. People choose villages for community and for children, I think, and remote work has made that possible for far more of them. Can I pick up the family option? Living near parents is often about childcare and about looking after them later, and it’s less of a free choice than the others.
B: True. And moving abroad? I did it for two years, and the first year was harder than anyone tells you.
A: In what way, if you don’t mind me asking?
B: Everything takes twice as long: the bank, the doctor, even making friends.
A: That’s interesting, because it links to adaptation, which is the decision question. Before that, the housing estate: I’d argue people choose it by elimination, it’s what they can afford with a garden.
B: Which makes it the least romantic but the most common. So, hardest to adapt to?
A: I’d say another country, for the reasons you gave, with the village a close second if you’ve only ever lived in a city. Would you go along with that?
B: Yes, another country, without much doubt.
C1-06: Reducing stress among university students (2 plus 1 minutes)
C1 Part 3. Timer: 15 seconds to read, 2 minutes to discuss, 1 minute to decide.
Interlocutor: Now, I’d like you to talk about something together for about two minutes. A university wants to reduce stress among its students. Here are some ideas it is considering, and a question for you to discuss. First you have some time to look at the task. (15 seconds)
Written task: central question “How effective might these ideas be in reducing student stress?”; prompts: longer deadlines for assignments / free counselling / better sports facilities / fewer exams and more coursework / mentoring by older students.
Now, talk to each other about how effective these ideas might be in reducing student stress.
After two minutes: Thank you. Now you have about a minute to decide which two ideas the university should introduce first.
Partner side (for a study partner or the AI to play): Candidate B’s likely lines: doubts longer deadlines help (“people just start later”); strongly for counselling but worries about waiting lists; thinks sport is underrated; asks A whether coursework is really less stressful; in the decision minute wants counselling and mentoring.
What this task really tests: Tests Interactive Communication and Discourse Management: the candidate evaluates each idea against the printed question (effectiveness), turns a prompt into a genuine question for the partner, and reasons for the pair of decisions rather than naming two favourites. The official C1 lesson plan for Speaking Parts 3 and 4 tells candidates “it is very important to interact with the other candidate in Part 3. You will get marks for this”, and that running out of time before all the prompts or a final decision is fine “as long as you discuss the topic together”.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
A: Let’s begin with deadlines, because I suspect it’s the least effective. Give people an extra week and most of them start a week later. Would you agree, or am I being cynical?
B: No, that’s been my experience too. Counselling is different, though; that addresses the cause.
A: It does, provided the waiting list isn’t six weeks, which is the problem at a lot of universities. So the effectiveness depends entirely on funding. What about sport? I think it’s underrated, because exercise is one of the few things with solid evidence behind it.
B: Yes, although the students who are most stressed are often the ones who feel they can’t spare an hour.
A: Good point, which brings us to exams versus coursework. Is coursework actually less stressful, or is it just stress spread thinly across the whole term?
B: I’d say spread thinly, and it rewards people who are organised.
A: Right, so it moves the stress rather than removing it. And mentoring: I like this one because it’s cheap and because a second-year telling you “I felt exactly like this” does something a counsellor can’t.
B: Agreed. So, two to introduce first?
A: Counselling for the people in real difficulty, and mentoring for everyone else, because together they cover both ends. Unless you’d swap mentoring for sport?
B: No, I’d keep mentoring. Counselling and mentoring.
C1-07: Why people change careers (2 plus 1 minutes)
C1 Part 3. Timer: 15 seconds to read, 2 minutes to discuss, 1 minute to decide.
Interlocutor: Now, I’d like you to talk about something together for about two minutes. Here are some reasons people give for changing career, and a question for you to discuss. First you have some time to look at the task. (15 seconds)
Written task: central question “How might these factors influence someone’s decision to change career?”; prompts: better pay / boredom / new technology replacing jobs / family responsibilities / health.
Now, talk to each other about how these factors might influence someone’s decision to change career.
After two minutes: Thank you. Now you have about a minute to decide which factor would be the most difficult to ignore.
Partner side (for a study partner or the AI to play): Candidate B’s likely lines: thinks pay is overrated as a motive once basic needs are met; has a parent who retrained because of technology; asks “Isn’t boredom a luxury?”; in the decision minute argues for health, then accepts family if A makes the case.
What this task really tests: Tests Interactive Communication: the candidate reacts to a personal story with a question instead of moving straight to the next prompt, then folds the story back into the task. That is what the Band 5 phrase “widens the scope of the interaction” sounds like. Also note the distinction between pushing and stopping a decision, which answers the printed question precisely.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
A: Can I start with pay, since it’s the one everyone mentions first? My feeling is that it pushes people to change jobs rather than careers; you don’t retrain as a nurse for the money. Does that match your view?
B: Largely, yes. Once you can pay the rent, pay matters less than people think.
A: Boredom is more interesting. It’s a slow pressure, and I think it drives more career changes than we admit, usually around forty. Although you could say boredom is a luxury.
B: I was about to say exactly that. My father didn’t have the luxury; his job disappeared when the factory automated.
A: That’s the technology prompt in one sentence. In that case the change wasn’t a decision at all, it was forced, which is a different kind of influence. How did he cope with it?
B: Badly at first, then he retrained in logistics and actually prefers it.
A: Which suggests the forced changes sometimes work out. Family responsibilities, I’d say, are the opposite: they stop people changing career, because the risk feels irresponsible when children depend on you. And health can go either way; a back injury ends one career and a scare starts another.
B: So, hardest to ignore?
A: I’d say health, because it removes the choice, but you could argue technology does the same. Which would you pick?
B: Health. Technology you can sometimes see coming; health you can’t.
C1-08: A company reducing its environmental impact (2 plus 1 minutes)
C1 Part 3. Timer: 15 seconds to read, 2 minutes to discuss, 1 minute to decide.
Interlocutor: Now, I’d like you to talk about something together for about two minutes. A large company wants to reduce its impact on the environment. Here are some measures it is considering, and a question for you to discuss. First you have some time to look at the task. (15 seconds)
Written task: central question “How much difference might these measures make?”; prompts: allowing staff to work from home / banning short-haul business flights / a plant-based staff canteen / switching to electric company vehicles / a paperless office.
Now, talk to each other about how much difference these measures might make.
After two minutes: Thank you. Now you have about a minute to decide which measure employees would be most likely to resist.
Partner side (for a study partner or the AI to play): Candidate B’s likely lines: thinks paperless is symbolic; supports the flight ban; worries working from home just moves energy use to homes; asks “Would you accept a plant-based canteen?”; in the decision minute says the canteen would be resisted most, then considers the flight ban.
What this task really tests: Tests Discourse Management in dialogue: proposing a way to organise the prompts (“sort these into big and small”) is a C1 move that gives both candidates a structure, and the printed question (how much difference) is kept in view with words like measurable, caveat, symbolic. Decision reached by argument about who is affected, not by preference.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
A: Shall we start by sorting these into big and small? My instinct is that the paperless office is almost symbolic; the carbon in paper is tiny compared with flights or vehicles. Would you put it at the bottom too?
B: I would. The flight ban is the opposite: one return flight to a meeting is worth months of printing.
A: Agreed, and it’s measurable, which companies like. Working from home is less clear-cut. The commute disappears, but you’re heating five hundred homes instead of one building, so the saving depends on the season and the country.
B: That’s a fair caveat. And the electric vehicles?
A: Real, but slow and expensive, and it only helps if the electricity is clean. Can we talk about the canteen, because that’s the one that touches people personally? The emissions difference is genuine, but it’s also the measure most likely to feel like the company telling you what to eat.
B: Exactly. Would you accept it yourself?
A: Two days a week, happily; five days, I’d bring a sandwich, which rather defeats the point.
B: So, which would employees resist most?
A: The canteen, I think, precisely because it’s personal and daily, whereas a flight ban only affects the people who fly. Unless you think frequent flyers would fight harder?
B: They might fight harder, but there are fewer of them. The canteen.
C1-09: Discussion: pressure, education and responsibility (5 minutes)
C1 Part 4. Timer: 5 minutes for both candidates together.
Interlocutor (after the student stress task, questions used in order as appropriate):
Some people say that a certain amount of stress is good for students. What do you think?
Do you think universities should be responsible for their students’ mental health, or is that the job of families and doctors?
How far do you agree that social media has made young people more anxious?
Is it better to prepare students for a competitive world or to protect them from it?
Do you think employers put too much emphasis on exam results?
Partner side (for a study partner or the AI to play): The interlocutor may address a question to one candidate and then ask the other “What do you think?” or “Do you agree?” Candidate B’s likely positions: a little stress sharpens people but chronic stress does not; universities have a duty of care because they created the environment; thinks the social media link is exaggerated.
What this task really tests: Part 4 at C1 gives Lexical Resource, Grammatical Resource and Interactive Communication their hardest test, because the C1 Band 5 Global Achievement descriptor covers “a wide range of topics, including unfamiliar and abstract ones”. The strong answer makes distinctions (pressure versus stress), extends the partner’s point rather than repeating it, and hedges honestly (“the research I’ve seen is mixed”). The handbook’s Part 3 advice, which applies here too, warns that “simply agreeing or disagreeing with their partner, or echoing what their partner has said” will not show what you can do.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
I’d distinguish between pressure and stress. A deadline that makes you focus for a week is pressure, and most people perform better with some of it. Stress, in the sense of lying awake for a month, doesn’t improve anybody’s work; it just makes them ill. So the claim is half true, and the danger is that institutions use the true half to excuse the other.
(After B says universities created the environment.) I’d go further than you, actually. It isn’t only that they created the environment; they also profit from it, through fees. If you charge someone nine thousand a year and design a system with five exams in two weeks, you can’t then say their anxiety is a matter for their GP. Families and doctors matter, of course, but the university controls the causes.
On social media, I’m more cautious than most people. Some researchers argue that anxiety among young people was already rising before smartphones, and the research I’ve seen is mixed. What I would say is that it removes the pause: a bad exam result used to stay in your room; now it’s compared with forty other people’s before you’ve left the building.
C1-10: Discussion: companies, individuals and the environment (5 minutes)
C1 Part 4. Timer: 5 minutes for both candidates together.
Interlocutor (after the company environmental measures task):
Who do you think has more responsibility for protecting the environment, companies or individuals? (Why?)
Some people say that measures like these are more about a company’s image than about the environment. How far do you agree?
Do you think governments should force companies to change, or let them decide for themselves?
Would you turn down a well-paid job at a company that harmed the environment?
Are people in your country generally willing to change their habits for environmental reasons?
Partner side (for a study partner or the AI to play): Follow-up prompts: “What do you think?”, “Do you agree?”, “Why?” Candidate B’s likely positions: companies, because scale matters; thinks image and substance are not mutually exclusive; would take the job and try to change things from inside.
What this task really tests: Tests Interactive Communication and Discourse Management: referring back to the Part 3 task (“the paperless office we discussed earlier”), building on the partner’s idea with a test rather than a repeat, and ending with self-aware honesty rather than a slogan. The interlocutor’s Global Achievement mark rewards exactly this kind of sustained, natural conversation on an abstract topic.
Show the model performance (try the task first)
Model performance, written by TalkDrill (not a real candidate)
I’d put more responsibility on companies, simply because of scale. If one person stops flying, it changes nothing measurable; if one airline changes its fleet, it does. That said, companies respond to customers, so individual choices aren’t pointless, they’re just indirect. The honest answer is that the responsibility is shared, but unequally.
On image versus substance, I’d agree with you that the two aren’t mutually exclusive, but I’d add a test. If the measure costs the company something real, a flight ban that slows down deals, for instance, it’s probably sincere. If it costs nothing and appears on the front of the annual report, it’s probably marketing. The paperless office we discussed earlier is a good example of the second kind.
Would I turn down the job? I’d like to say yes, but I think it depends on the alternatives and the degree of harm. An oil company, probably yes. A clothing retailer with a poor record, I might take it and push for change, as you suggested, although I’m aware that’s exactly what everyone tells themselves.
Six full mock tests built from the bank
A part on its own is practice; a full test at the real length is a mock. Each mock below chains bank tasks in the official order, with the Part 1 and Part 4 questions the bank does not cover written here by TalkDrill. Run each with one timer for the whole test and do not stop between parts. Candidate A takes the first Part 2 task, Candidate B the second.
| Mock | Part 1 | Part 2 (A, then B) | Part 3 | Part 4 | Length |
|---|---|---|---|---|---|
| B1 mock 1 | B1 questions set 1 below | B1-01, B1-02 | B1-06 | B1 questions set 1 below | About 12 minutes |
| B1 mock 2 | B1 questions set 2 below | B1-03, B1-04 | B1-08 | B1 questions set 2 below | About 12 minutes |
| B2 mock 1 | B2-01 | B2-03, B2-04 | B2-06 | B2-09 | 14 minutes |
| B2 mock 2 | B2-02 | B2-05, B2-03 (roles swapped from mock 1) | B2-08 | B2-10 | 14 minutes |
| C1 mock 1 | C1 questions set 1 below | C1-01, C1-02 | C1-06 | C1-09 | 15 minutes |
| C1 mock 2 | C1 questions set 2 below | C1-03, C1-04 | C1-08 | C1-10 | 15 minutes |
Spare tasks for extra Part 3 practice: B1-07, B1-09, B1-10, B2-07, C1-05 and C1-07. Spare B1 Part 2 photograph: B1-05.
Part 1 and Part 4 questions for the mocks (written by TalkDrill)
B1 set 1
Part 1 (to each candidate in turn): What's your name? Where do you live? Do you work or are you a student? What do you enjoy doing at the weekend? Tell us about your favourite place in your town. How often do you use English outside the classroom?
Part 4 (after B1-06, to both): Do you prefer visiting cities or the countryside? (Why?) What do you like doing when you visit a new place? Is it better to travel with friends or with family? (Why?) What's the most interesting place you have visited? Would you like to live in a big city in the future? (Why? / Why not?)
B1 set 2
Part 1: What's your name? Where do you live? Do you work or are you a student? How do you usually travel to school or work? What did you do last weekend? Do you prefer shopping alone or with friends? (Why?)
Part 4 (after B1-08): Have you ever helped to raise money for something? Do you enjoy school trips? (Why? / Why not?) Is it better to save money or to spend it? (Why?) What do young people in your country usually spend their money on? Would you like a part-time job? (Why? / Why not?)
C1 set 1
Part 1 (to each candidate in turn, after names and where you are from): How long have you been studying English, and what has kept you going? What's something you've learned recently outside your studies or work? How do you usually switch off after a demanding day? If you could spend a year living anywhere in the world, where would you choose? (Why?)
C1 set 2
Part 1: What does a typical week look like for you at the moment? Is there a skill you wish you had learned earlier? (Why?) How important is it to you to plan your free time? (Why? / Why not?) Which teacher or colleague has taught you the most? (Why?)
The interlocutor in the real test chooses from a frame of questions, so do not prepare answers to these. Use them to practise answering in two or three sentences with a reason: the B1 handbook says candidates "should avoid giving one-word answers" but "are not expected to give very long answers at this stage".
Score yourself on the official criteria
Listen to each mock recording once per criterion. The criteria and band wording are Cambridge's, from the handbooks' assessment scales; the playback questions are ours. Each analytical criterion is marked on bands 0 to 5, and Band 4 and Band 2 are defined as sharing features of the bands either side. Cambridge's published examiner commentaries also use half marks: in the B2 First commentary for Giulia, Tabatha and Elena, Tabatha received 3.5 for Grammar and Vocabulary and 4 for Interactive Communication.
| Criterion | B2 First Band 5 wording (handbook) | Ask yourself on playback |
|---|---|---|
| Grammar and Vocabulary (at C1: Grammatical Resource and Lexical Resource, marked separately) | "Shows a good degree of control of a range of simple and some complex grammatical forms. Uses a range of appropriate vocabulary to give and exchange views on a wide range of familiar topics." | Did I use at least one complex form per turn (a conditional, a relative clause, a modal perfect)? Did I repeat "good", "nice" and "interesting"? |
| Discourse Management | "Produces extended stretches of language with very little hesitation. Contributions are relevant and there is a clear organisation of ideas. Uses a range of cohesive devices and discourse markers." | In Part 2, did I reach the printed question by 25 seconds? Did I link with "whereas", "as for", "on the other hand", or just "and"? |
| Pronunciation | "Is intelligible. Intonation is appropriate. Sentence and word stress is accurately placed. Individual sounds are articulated clearly." | Could a stranger write down every word? Did my voice fall at the end of statements? Did I stress the contrasting words in comparisons? |
| Interactive Communication | "Initiates and responds appropriately, linking contributions to those of other speakers. Maintains and develops the interaction and negotiates towards an outcome." | How many questions did I ask my partner in Part 3? Did I react to their point before adding mine? Did I propose the decision as a question? |
| Global Achievement (the interlocutor's mark) | "Handles communication on a range of familiar topics, with very little hesitation. Uses accurate and appropriate linguistic resources to express ideas and produce extended discourse that is generally coherent." | Across the whole test, did I keep going without long silences, and did it all make sense to someone hearing it once? |
The B1 descriptors use the same four headings at a lower level (Band 5 Interactive Communication at B1: "Maintains and develops the interaction and negotiates towards an outcome with very little support"). The C1 Band 5 Interactive Communication descriptor adds "Interacts with ease" and "Widens the scope of the interaction". Every band, with what each sounds like, is on the criteria and scale page.
Official material to use alongside
TalkDrill's tasks are practice in the official format. Cambridge's own material is free and you should use it too, especially the Speaking test videos, which show the format and the examiners' scripted wording better than any description.
| What | Where | Why it helps |
|---|---|---|
| Speaking test videos with examiner comments, B2 First: Giulia, Tabatha and Elena; Julia and Isabela; Natalia and Enzo | The B2 First preparation page | Real candidates, real timing, and a PDF of examiner comments with scores for each criterion. The Giulia, Tabatha and Elena test is a group of three. |
| Speaking test videos with examiner comments, C1 Advanced: Marta, Giovanna and Anna; Pedro and Isabela; Teresa and Antonio | The C1 Advanced preparation page | Shows what C1 speculation and Part 3 negotiation sound like at the level. |
| Sample papers, including the Speaking tasks | The same preparation pages, in the paper-based sample test folders | Official photo sets and Part 3 task sheets to run with the timers on this page. |
| Handbooks for teachers | The B2 First handbook (PDF); the C1 Advanced handbook is listed on the C1 preparation page | The interlocutor frame, the advice by task and the full assessment scales. |
| Lesson plan for C1 Advanced Speaking Parts 3 and 4 | Cambridge PDF (UCLES 2021) | A model decision-minute exchange and a table of interaction phrases, used on our Part 3 page. |
The format and timings on this page were checked against the B2 First exam-format page on 21 October 2026.
A four-week route through the bank
About 20 minutes a day, five days a week. Each day is one task, recorded, scored on one criterion. If your Speaking test is closer than four weeks, run one full mock today, find the part where your score sheet looks worst, and start at that week.
| Week | Focus | B1 Preliminary | B2 First | C1 Advanced |
|---|---|---|---|---|
| 1 | Part 2 on your own, timed to the second | B1-01 to B1-05, one a day | B2-01 and B2-02 (Part 1), then B2-03 to B2-05 | C1 Part 1 sets, then C1-01 to C1-04 |
| 2 | Part 3 with a partner or the AI | B1-06 to B1-10 | B2-06 to B2-08, each twice with roles swapped | C1-05 to C1-08 |
| 3 | Part 4 and the first full mock | B1 mock 1, then the Part 4 questions on their own | B2-09, B2-10, then B2 mock 1 | C1-09, C1-10, then C1 mock 1 |
| 4 | Second mock, review, rest | B1 mock 2; compare with week 3 | B2 mock 2; compare with week 3 | C1 mock 2; compare with week 3 |
In week 4, play your week 1 and week 4 recordings of the same task back to back. That comparison is the most useful ten minutes of the month. For the phrases each part needs, use the Part 2 phrase bank and the Part 3 phrase bank; for a full study plan around the written papers, see the eight-week B2 First roadmap. If you are sitting a paper-based exam, the remaining 2026 dates are B2 First on 7 November, 24 November and 12 December, and C1 Advanced on 14, 25 and 28 November and 5 and 12 December, per the official 2026 dates PDF; your centre will tell you the Speaking date, which can be a different day: the official dates page says Speaking test dates "are subject to local provision". All dates are on the dates page, and everything else on the Cambridge English hub.