The B2 First Speaking test in one paragraph
The B2 First (still called FCE by most teachers) Speaking test is 14 minutes for a pair of candidates, 20 minutes for a group of three, and has four parts, per the official B2 First exam-format page. Two examiners sit with you: the interlocutor asks the questions and hands you the booklet, and the assessor listens and writes marks. Part 1 is a 2-minute interview about you. Part 2 gives each candidate two photographs and 1 minute to compare them, then a 30-second question to the partner. Part 3 is a 2-minute discussion with your partner around five written prompts, followed by 1 minute to reach a decision. Part 4 is a 4-minute discussion led by the examiner on the same theme. You are marked on four criteria (Grammar and Vocabulary, Discourse Management, Pronunciation, Interactive Communication) by the assessor and on Global Achievement by the interlocutor, and the handbook is explicit that you are assessed "on your own individual performance and not in relation to each other". The one thing most candidates get wrong: they prepare speeches. The test rewards people who listen, respond and keep a conversation moving, and Part 3 is the only part built entirely around talking to your partner, so it is where the Interactive Communication mark is easiest to earn or to lose.
- 14 minutes per pair; 20 minutes per group of three. Same format for paper and digital B2 First, and for B2 First for Schools.
- Two examiners. One talks to you and "could be online, examining remotely"; the other only listens. Examiners may enter marks on a phone app.
- Part 1 interview 2 minutes; Part 2 long turn 1 minute each plus a 30-second response (4 minutes in total); Part 3 2-minute discussion plus 1-minute decision (4 minutes in total); Part 4 discussion 4 minutes.
- Marked on Grammar and Vocabulary, Discourse Management, Pronunciation, Interactive Communication and Global Achievement, each on a 0 to 5 band.
- Speaking is one of five scores averaged into the overall Cambridge English Scale result; 160 is the bottom of Grade C, 180 the bottom of Grade A (results page).
| Part | Official timing (pair) | Who you talk to | What the examiners are listening for (handbook "focus") |
|---|---|---|---|
| 1 Interview | 2 minutes | The interlocutor | General interactional and social language |
| 2 Long turn | 1 minute each, plus a 30-second response; 4 minutes in total | On your own, then a short reply | Organising a larger unit of discourse; comparing, describing, expressing opinions |
| 3 Collaborative task | 2 minutes discussion, then 1 minute to decide; 4 minutes in total | Your partner | Sustaining an interaction; exchanging ideas, justifying, agreeing and disagreeing, negotiating |
| 4 Discussion | 4 minutes | The interlocutor and your partner | Expressing and justifying opinions, agreeing and disagreeing, speculating |
Part 1: the interview (2 minutes)
The interlocutor works from a scripted frame. The B2 First handbook prints it: a greeting, "My name is ... and this is my colleague ...", "Can I have your mark sheets, please?", then "Where do you live?" to each candidate, then "First we'd like to know something about you" and one or more questions chosen from categories such as likes and dislikes, special occasions and media. Each question has follow-ups in brackets, usually "Why?" or "What do you do?". The handbook says the questions "relate to the candidates' own lives and focus on areas such as work, leisure time and future plans", and that candidates "are not actively invited to talk to each other in this part".
Ten Part 1 questions in the official style (written by TalkDrill)
- How do you usually spend your weekends? (Why?)
- Do you prefer studying in the morning or in the evening? (Why?)
- Tell us about a place in your town you like going to.
- Do you enjoy cooking? (What sort of things do you cook?)
- What did you do on your last holiday?
- Are you planning to do anything special next weekend? (What?)
- How much time do you spend online in a day? (Would you like to spend less?)
- Tell us about a series or film you have watched recently.
- Do you prefer to celebrate your birthday with family or with friends? (Why?)
- Is there a sport you would like to try? (Why that one?)
What a passing answer and a strong answer sound like
Both answers below are written by TalkDrill. The labels are our reading of the B2 First Speaking scales in the handbook, where Band 3 is the B2 standard and Band 5 is the top; they are not Cambridge marks.
Question: How do you usually spend your weekends?
Passing (Band 3 features): "Normally I spend my weekends with my friends. On Saturday we go to the city centre and we have a coffee, and on Sunday I stay at home with my family because we have lunch together. Sometimes I study if I have an exam." Relevant, extended beyond a phrase, accurate simple grammar, a couple of linkers. Enough.
Strong (Band 5 features): "It depends on the week, to be honest. If I've had a lot of classes, I'll probably sleep in on Saturday and then meet friends in the afternoon, usually somewhere in the centre where we can sit outside. Sundays are quieter; we tend to have a long family lunch, which I've come to appreciate more since my brother moved away." A conditional, a future form, "tend to", "have come to appreciate", a relative clause, and a reason the listener did not ask for but is glad to hear. Still under 20 seconds.
Where Part 1 marks go
- One-word answers. "Yes." "With friends." The interlocutor will prompt "Why?", and every prompt is support, while Interactive Communication at Band 3 asks for interaction "with very little support".
- Rehearsed monologues. The handbook warns that "rehearsed speeches should be avoided as these might be inappropriate for the question asked". A memorised paragraph about your hometown delivered after "Do you like cooking?" is heard for what it is.
- Answering your partner's question. In Part 1 the questions go to one candidate at a time. Listen, and wait.
- Present simple for everything. Spanish, Italian and Portuguese speakers often say "I live here since five years". Part 1 is the cheapest place to show a present perfect, a past, and a future form. Our Romance-language page drills the habit.
Part 2: the long turn with two photos (1 minute plus 30 seconds)
Per the format page, "the examiner gives the student a pair of photographs to talk about and they have to speak for 1 minute without interruption. The questions about the photographs are written at the top of the page." After you finish, your partner "has to answer a short question from the examiner about their photographs", for about 30 seconds. Then the roles swap with a new pair of photos. The handbook adds the detail that decides the mark: candidates "are expected to point out similarities and differences between the photographs and then move on to deal with the question, answering it with reference to both photographs", and "are not expected to give detailed descriptions of each picture".
Two photo sets in the official style (TalkDrill)
| Set | Photo A | Photo B | Printed question | 30-second question to the partner |
|---|---|---|---|---|
| 1 Learning | Four teenagers around a table in a library with laptops and notes | An older man alone at a kitchen table with headphones and a language app | Why might the people have chosen to learn in these ways? | Which way of learning do you prefer? (Why?) |
| 2 Celebrations | A large family meal outdoors with a long table and children running | Two friends toasting with coffee in a small cafe | What might the people be enjoying about celebrating in these ways? | Do you prefer big celebrations or small ones? (Why?) |
A strong 1-minute answer, set 1 (TalkDrill, about 150 words)
"Both pictures show people learning, but in very different settings. In the first one, a group of teenagers are studying together in what looks like a school library, with laptops and quite a lot of paper, whereas in the second a man is on his own at home, using an app with headphones on. The first group have probably chosen to work together because they're preparing for the same exam and can explain things to each other; it also looks more sociable, which matters at that age. The man, on the other hand, might have chosen the app because he's fitting learning around work or family, so he needs something he can do for twenty minutes in the kitchen. He may also prefer to make mistakes in private. So in both cases the method seems to suit the person's situation rather than being the best way in itself."
Notice the shape: one sentence to open, two to compare, then most of the minute on the printed question, with "probably", "might", "may" doing the speculating, and a closing sentence that answers the question for both photos. A passing answer tends to spend 40 seconds describing ("there are four people, they have laptops, there are books") and then rushes the question. The full phrase bank and a 10-second-per-move timeline are on our Part 2 phrases page.
The 30-second response: short, but marked
When your partner finishes, the interlocutor turns to you with a short question. Thirty seconds is three or four sentences: an answer, a reason, a small example. "I'd say the second way, the app, because I'm quite impatient and I like to go at my own pace. Having said that, I did pass my last exam by studying with two friends, so maybe I need both." "The first one" with nothing after it gives the assessor nothing to mark.
Where Part 2 marks go
- Describing instead of comparing. The task says compare. Say one similarity, two differences, then answer the question.
- Ignoring the printed question. It is written above the photos so you cannot forget it. Candidates who never answer it lose on Discourse Management ("contributions are relevant").
- Stopping at 35 seconds. The handbook tells teachers to time practice "so that your students get a feel for how long a minute is". Record yourself; a minute is longer than it feels.
- Talking during your partner's turn. The handbook: "They should not speak during their partner's long turn."
Part 3: the collaborative task (2 minutes plus 1 minute)
This is the part the test is built around. The interlocutor reads: "Now, I'd like you to talk about something together for about two minutes." You see a written question in the middle of the page with five prompts around it, and you get 15 seconds to look at it before "Now, talk to each other about ...". After two minutes: "Now you have about a minute to decide which ...". 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", that you "are not penalised if they fail to reach a negotiated decision", and that you are assessed on your "ability to hold a conversation, to turn-take appropriately, and to use the language of negotiation and collaboration".
Two Part 3 tasks in the official layout (TalkDrill)
Task A
Question in the centre: Why do some people find it hard to save money? Prompts: eating out · online shopping · rent and bills · holidays · following friends. Decision question: Which of these is the hardest habit to change?
Task B
Question in the centre: Is it a good idea for teenagers to have a part-time job? Prompts: managing time · learning about money · meeting people · school results · independence from parents. Decision question: Which of these matters most to parents?
Forty seconds of a strong exchange, Task A (TalkDrill)
A: "Shall we start with eating out? I think it's the one people underestimate, because each meal seems cheap." B: "True, but for me online shopping is worse, since it's so easy to buy something at midnight without thinking. Don't you find that?" A: "Yes, and the apps are designed for it. Although I'd say rent and bills are different, because you can't really choose them." B: "Good point, so maybe that one doesn't belong with the others. What about holidays? People book them even when they can't afford them." A: "Partly because of friends, I think, which brings in the last prompt ..."
Count what is happening: an invitation ("Shall we start with"), agreement with a development ("True, but"), a question back ("Don't you find that?"), a polite disagreement ("Although I'd say"), a link between prompts. That is Band 5 Interactive Communication in the handbook's words: "initiates and responds appropriately, linking contributions to those of other speakers". A passing pair takes the prompts in turn, one each, with "I agree" between them; it is relevant and it is enough for Band 3, but nothing in it links to the partner. The turn-taking language, and why a 60-second monologue costs marks, is on our Part 3 phrases page.
Where Part 3 marks go
- The monologue. One candidate covers all five prompts in 70 seconds. The other has nothing left. Both lose on Interactive Communication, the first for not inviting, the second for not interrupting politely.
- Deciding too early. Agreeing in the first 30 seconds leaves 90 seconds of silence. The decision comes after the interlocutor asks for it.
- Reading the prompts aloud. "Eating out. Online shopping." is not discussion. Turn each prompt into an opinion with a reason.
- Forgetting the question in the middle. The prompts are ideas for answering it; the handbook says you should "answer the question by exploring the different prompts".
Part 4: the discussion (4 minutes)
The interlocutor now asks questions "which encourage the candidates to discuss further and broaden the topics introduced in Part 3", and the handbook draws the line that matters: these questions "ask primarily for an evaluation rather than for information". Questions may go to one candidate or to both, and the interlocutor will use "What do you think?", "Do you agree?" and "And you?" to bring the second candidate in. The handbook also says that "the question Why? is useful for nearly all Part 4 questions and the interlocutor will often ask this question if students fail to give more than a minimal response", and that candidates "will not be judged on their opinions, only on the language they use to express their opinions".
Ten Part 4 questions following Task A and Task B (TalkDrill)
- Do you think schools should teach students how to manage money? (Why? / Why not?)
- Some people say it is harder to save money today than it was for our parents. What is your opinion?
- Is it better to spend money on experiences or on things? (Why?)
- Do you think advertising makes people buy things they do not need? (Why? / Why not?)
- How important is it for young people to be financially independent? (Why?)
- Do you think teenagers who work are more responsible than those who do not? (Why? / Why not?)
- Should parents pay their children for helping at home? (Why? / Why not?)
- Is it a good thing that many students now work while they study at university? (Why? / Why not?)
- Do you think people worry too much about money? (Why? / Why not?)
- In the future, do you think people will use cash at all? (Why? / Why not?)
Passing and strong, question 2 (TalkDrill)
Passing: "Yes, I think it's harder now because everything is more expensive. For example, the rent in my city is very high, so young people can't save. Also we have more things to buy, like phones." Clear, relevant, a reason and an example; Band 3 on Discourse Management.
Strong: "In some ways, yes. Housing takes a much bigger share of a salary than it used to, at least where I live, so even people with decent jobs struggle to put anything aside. But I'm not sure it's only about prices. We're also exposed to far more things to want; my grandmother wasn't shown fifty products a day on her phone. So I'd say it's harder, but partly for reasons we could control if we chose to." A concession, a contrast, a speculation about cause, and a conclusion that evaluates rather than reports. This is what "handles communication on a range of familiar topics, with very little hesitation" sounds like in the Band 5 Global Achievement descriptor.
Where Part 4 marks go
- Information instead of evaluation. "In my country people use cards" answers a Part 1 question. Part 4 wants "whether that is a good thing, and why".
- Not bringing your partner in. "What do you think, Marta?" after your answer is Interactive Communication you can earn in Part 4 as well as Part 3.
- Fading in the last minute. The test is 14 minutes, and it is easy to run out of ideas by minute 12. Practise full-length mocks, not single parts.
How the 14 minutes are marked
Per the handbook, the assessor marks four analytical criteria using band descriptors from 0 to 5, and the interlocutor gives a single Global Achievement mark on the same 0 to 5 scale; Bands 2 and 4 are defined as performances that share features of the bands either side. Assessment "is based on performance across all parts of the test", so a weak Part 1 can be recovered in Parts 3 and 4. The table gives the Band 3 descriptor (the B2 standard) and the Band 5 descriptor (the top), lightly shortened from the handbook, with our note on what to practise.
| Criterion | Band 3 (handbook) | Band 5 (handbook) | What to practise (TalkDrill) |
|---|---|---|---|
| Grammar and Vocabulary | Good control of simple forms, attempts some complex forms; a range of vocabulary on a range of familiar topics | Good control of a range of simple and some complex forms; a range of vocabulary on a wide range of familiar topics | Conditionals, relative clauses, passives and modals said aloud until they come without effort |
| Discourse Management | Extended stretches despite some hesitation; relevant, very little repetition; a range of cohesive devices | Extended stretches with very little hesitation; clear organisation of ideas; cohesive devices and discourse markers | One-minute answers with an opening, a comparison and an answer to the question |
| Pronunciation | Intelligible; intonation generally appropriate; stress generally accurate; sounds generally clear | Intelligible; intonation appropriate; sentence and word stress accurately placed; sounds clear | Word stress and intonation, not accent; see the word stress guide |
| Interactive Communication | Initiates and responds appropriately; maintains and develops the interaction with very little support | Initiates and responds, linking contributions to those of other speakers; negotiates towards an outcome | Questions back, polite disagreement, building on what your partner said |
The Speaking mark becomes a Cambridge English Scale score for Speaking, and per the official results page the overall result is the average of the five scores: 160 to 172 is Grade C, 173 to 179 Grade B, 180 to 190 Grade A with a certificate stating C1, 140 to 159 Level B1.
How the five marks add up (official Scale guide)
Cambridge's guide for teachers, The Cambridge English Scale explained (November 2023), states the B2 First weighting: each of the assessor's four marks is doubled, the interlocutor's Global Achievement mark is multiplied by four, and examiners may award half marks, so Speaking is out of 60. For official practice tests only, it places 36 marks at 160 (the bottom of B2), 54 at 180 (C1) and 24 at 140 (B1), and it warns that the tables "should not be used to try to predict precise scores in the live exam". Two things follow (our arithmetic). Global Achievement is 20 of the 60 marks, a third of the test, from the examiner you talk to. And 36 of 60 is an average of Band 3 across the board: the B2 standard really is Band 3.
An official example: Natalia and Enzo
Cambridge publishes a B2 First Speaking video of two candidates, Natalia and Enzo, with examiner comments and marks. Natalia received 3.5 for Grammar and Vocabulary, 3.5 for Discourse Management, 4 for Pronunciation, 4 for Interactive Communication and 4 for Global Achievement. Enzo received 4, 4, 4.5, 4 and 4. The examiners note that Natalia's Part 2 "is rather short at only 36 seconds rather than a full minute", and that Enzo "does little to negotiate towards an outcome with his partner" in Part 3.
Our arithmetic, not a Cambridge figure: with the guide's weighting Natalia's marks come to 46 of 60 and Enzo's to 49 of 60, both between the 36 that the guide places at 160 and the 54 it places at 180. Cambridge gives no Scale score for these videos. The point is where the marks went: a short long turn and thin negotiation, exactly the losses listed under Parts 2 and 3 above.
The full descriptors for every band, the Global Achievement scale at all three levels and the conversion tables for B1 Preliminary and C1 Advanced are on our criteria and scale page.
A practice route to test day
The dates. Per the 2026 examination dates PDF (version 4, June 2026) the remaining paper B2 First sittings are Saturday 7 November, Tuesday 24 November and Saturday 12 December 2026; the 2027 PDF (version 1) starts with Saturday 27 February 2027. B2 First Digital is "available throughout the year" through centres, and the PDFs tell you to check the Speaking date with the centre, because it is often a different day. Fees are set by each centre, not by Cambridge; our fees and registration page explains how to compare them, and the dates page lists every 2026 and 2027 sitting.
- Watch the official sample test once, with the handbook open. The B2 First preparation page links free sample papers, Speaking test videos with examiner comments and the handbook. Match what you see to the four parts above.
- Part 1 and Part 4 every day, five minutes. Answer three questions from the lists above out loud; one reason and one example each. Our AI interview practice asks follow-ups the way the interlocutor does.
- Part 2 timed, three times a week. Any two photos from your phone, a question you write above them, a 60-second timer. Listen back for "describe versus compare".
- Part 3 with a partner, or with the AI taking the second candidate's turns. This is the gap for most self-study candidates. Our paired practice bank runs the real timers: 15 seconds to read, 2 minutes, then 1 minute to decide.
- Two full 14-minute mocks in the last fortnight. Record them. Mark yourself against the four criteria, and check the first-language page for your language (Turkish and Vietnamese speakers have their own).
If you are deciding between B2 First and IELTS for a university application, the formats differ in a way that matters for Speaking: IELTS Speaking is one candidate with one examiner, B2 First is paired. Our IELTS Speaking hub covers the other side. The Cambridge English hub has the dates block and the Scale strip. If you are choosing a level, the C1 Advanced Speaking page shows the same four parts one level up, and the B1 Preliminary Speaking page one level down.