TOEFL iBT® Speaking questions and topics for the 2026 test

• Questions and time: 11 questions in about 8 minutes, according to ETS.
• Two tasks: Listen and Repeat (7 sentences) and Take an Interview (4 questions).
• Not adaptive: everyone who takes the same version of the test gets the same Speaking tasks.
• Each task starts with a short scenario that you both hear and read, with a picture of the setting or the speaker.
• No separate preparation time: you repeat each sentence after a short pause, and you have 45 seconds to answer each interview question. You plan while you speak.

You speak into a microphone, and your answers are recorded. See the full test format for all four sections.

What it tests: whether you can understand a sentence and say it back accurately and clearly.

What you hear: seven sentences from one situation on campus or in academic life, such as a guide or a manager explaining a place. A picture of the setting is on the screen. After each sentence there is a short pause, and then you repeat exactly what you heard. The sentences get longer and more complex as you go, and you have up to 8 to 12 seconds to record each one.

How to repeat a long sentence:

• Remember chunks of meaning, not single words. Sentence 3 has three chunks: [Students who have finished the safety course] [can use the workshop in the evening] [without a staff member present]. Three chunks are easier to hold than nineteen words.
• Copy the stress and rhythm. Say the important words, such as “safety course”, “evening” and “without”, a little more strongly, as the speaker did.
• Keep going. If you miss a word, say a close one and finish the sentence. Stopping halfway loses more than one wrong word.
• Protect the meaning. Dropping “not”, “without” or a time phrase changes what the sentence says.

A full set of seven sentences, broken into chunks with the stress marked, is in the full Listen and Repeat guide.

What it tests: whether you can answer questions clearly and in detail, at a natural pace, with accurate grammar and vocabulary.

What you hear: a recorded interviewer in a situation such as a research study or a scholarship interview. The interviewer asks four questions on one topic. The first questions are about facts and your own experience; the later ones ask for your opinion on a wider issue. You have 45 seconds for each answer.

A simple method: point, reason, example.

1. Point: answer the question directly in your first sentence.
2. Reason: say why, in one sentence.
3. Example: give one short, specific example from your life.
4. Close: finish with one sentence that links back to the question.

The answer gives the point in the first sentence, one reason, one specific example and a closing line. That takes about 45 seconds at a natural pace, so the speaker does not need to rush or fill silence.

For a plan for each of the four questions and a fully worked interview, read the full Take an Interview guide.

Use these original question sets to practice. Each set moves from your own experience to your opinion, like the real task. Set a timer for 45 seconds and answer aloud.

• Every answer is scored from 0 to 5 with ETS’s scoring guides.
  • Listen and Repeat: a 5 is a fully intelligible, exact repetition. Small changes that keep the meaning, such as a missing small word, score lower, and changes to the meaning score lower again.
  • Take an Interview: the guide looks at whether you answer the question with relevant, well-developed detail, keep a natural conversational pace, speak clearly with helpful rhythm and intonation, and use a range of accurate grammar and vocabulary.
• Who scores: ETS scores Speaking with its AI scoring engine, which is trained on human ratings. Human raters help check its accuracy.
• Your result: a Speaking band from 1 to 6, in half steps. See how the 1–6 band scale works, with the CEFR level for each band.

On Prepiva, your Speaking rating is an AI-generated practice estimate and is labeled provisional. How we estimate bands explains the method and its limits.

These traps come from Prepiva’s guides for the two Speaking tasks:

• Listen and Repeat: starting too slowly and running out of time on the long sentences; changing the meaning by dropping “not” or a time phrase.
• Take an Interview: long silences while you plan a perfect sentence; memorized answers that do not fit the question.

Prepiva’s free full practice test, Practice Exam 01, ends with a full Speaking section in the 2026 format. After a free sign-up, its score report includes an AI rating of your Speaking and Writing in that test. Take the full practice test to practice speaking under real timing.

For more Speaking practice with feedback, Prepiva Pro includes Speaking practice sets with AI evaluation, for $19.99 a month or $39.99 for 3 months. To prepare for the section before Speaking, see the Writing section, and for more listening practice, the Listening questions.