TOEFL iBT Speaking Section Overview: Tasks, Timing and Scoring

The Speaking section of the TOEFL iBT® test is now one of the shortest parts of the exam. It is also one of the easiest to underestimate. In about eight minutes you answer 11 items, with very little time to think between them.

This is a one-page reference for the section in the format ETS has used since 21 January 2026. All the numbers come from ETS’s Updated TOEFL iBT Test Overview, the TOEFL iBT 2026 test specifications and the Speaking Scoring Guide. If you are coming from the old format, read TOEFL Speaking in 2026: What Changed first.

Where Speaking sits in the test

ETS gives the sections in a fixed order: Reading, Listening, Writing, then Speaking. Speaking comes last, after more than an hour of concentration. ETS says the whole test takes under two hours, and there are no scheduled breaks.

That order matters for your preparation. You will speak when you are already tired. If you only ever practise speaking first thing in the morning with a fresh mind, you are practising for a different situation.

The section at a glance

Task type	What you do	Items	Time to respond	Scored on
Listen and Repeat	Hear a sentence once, then repeat it exactly	7	Up to 8–12 seconds per sentence	Accuracy of the repetition and intelligibility
Take an Interview	Answer a prerecorded interviewer’s questions about your experiences and opinions	4	45 seconds per question	Relevance and elaboration, pace, intelligibility, grammar and vocabulary

Total: 11 items, about 8 minutes. ETS’s estimate does not include time for things like adjusting the volume or reading the instructions.

How each task begins

ETS’s test specifications describe a common pattern. Each task starts with a scenario introduction, which you both hear and read on screen. It tells you where you are and why you are listening, for example “You are learning to welcome visitors to the zoo.” Visuals support the context. ETS mentions schematic images, maps, forms and photos of speakers.

For Listen and Repeat, the on-screen picture shows the setting, and your progress through the seven sentences follows related parts of the picture. For Take an Interview, you hear and see a scenario such as taking part in a research study, and then the interviewer starts asking questions.

Listen and Repeat in brief

• Seven sentences from one situation, such as a manager training you for a campus job.
  
• Each sentence is played once. After a short pause, you repeat exactly what you heard.
  
• The sentences get longer and more complex as you go, from short single-clause sentences to longer ones with dependent clauses, relative clauses and longer noun phrases.
  
• ETS says the sentences include natural spoken features such as contractions.
  
• Each response is scored from 0 to 5. A 5 means an exact, fully intelligible repetition.
  

We cover this task in depth in the next post.

Take an Interview in brief

• Four questions from one interviewer, all on one topic.
  
• 45 seconds per answer. ETS lists no separate preparation period, so plan to start speaking almost immediately.
  
• The early questions ask for facts and personal experience. The later ones ask you to state and support an opinion on a broader issue.
  
• Each response is scored from 0 to 5. A 5 is “a fully successful response” that is on topic, well elaborated, delivered at a natural pace, easy to understand, and uses a range of accurate grammar and vocabulary.
  

How the section is scored

Each of the 11 items has a maximum of 5 points, so the section has a raw maximum of 55. ETS converts that raw total to a Speaking band from 1 to 6 in half-band steps using a statistical linking procedure. The procedure is designed so that a given band means the same level of ability whichever version of the test you took.

ETS says Speaking and Writing responses are scored by its AI scoring engine, applying the criteria in the scoring guides. The engine is trained on human ratings, and human raters provide oversight to check accuracy.

Two practical points follow:

1. Listen and Repeat is more than half the items. Seven of 11 items is a lot of points. Neglecting this task is one of the most expensive mistakes you can make.
   
2. Every item counts on its own. A weak first interview answer does not ruin the section, so reset and keep going.
   

What you will hear

ETS says you may hear English accents from North America, the UK, Australia and New Zealand. Practise with a mix of voices so that an unfamiliar accent in sentence six doesn’t throw you.

A realistic pacing plan

Because there is no preparation time to manage, pacing in this section means managing attention:

• Before the section starts: take one slow breath and sit up straight. Your voice is clearer when you aren’t hunched over the keyboard.
  
• During Listen and Repeat: listen to the whole sentence before you start forming it. Don’t begin speaking in your head while the speaker is still talking.
  
• Between tasks: use the scenario introduction for the interview to picture the situation and the person you are talking to.
  
• During the interview: answer the question in your first sentence, then develop it. Watch the response clock, but don’t stop early just because you have made one point.
  

Practise this

Lingo Copilot Speaking runs the whole TOEFL speaking section in test order: one Listen and Repeat set followed by one interview set, with a response clock for every item. You get an AI estimate for each item on the 0–5 scale and an estimated Speaking band with its CEFR level for the full section.

Open TOEFL speaking practice, or create a free account. Your first full TOEFL speaking mock is free.

Scores in Lingo Copilot Speaking are AI estimates for practice, not official TOEFL scores. TOEFL and TOEFL iBT® are registered trademarks of ETS. Lingo Copilot Speaking is not affiliated with or endorsed by ETS.