TOEFL Speaking (2026): Master the Interview Format

In the 2026 format, TOEFL Speaking has two tasks. Listen and Repeat gives you a short recorded utterance and asks you to repeat it accurately — pronunciation, stress and rhythm all count. Take an Interview is the bigger change: a virtual interviewer asks you a question, listens to your answer, and then asks a natural follow-up. The follow-up can go in directions you did not plan for, which is the whole point.

The old independent Speaking tasks that rewarded a memorised 45-second monologue are gone. Everything is now closer to a real conversation.

The interview rewards spontaneous, natural answers. A memorised opening buys you no credit if your second sentence has to reply to something the interviewer just said. Candidates who trained heavily on template-based Speaking often score lower on the interview than in old-format practice — not because their English got worse, but because their strategy no longer matches the task.

The good news: the interview does not expect a perfect answer. It expects a real one. A relaxed, on-topic reply with some hesitation scores better than a polished-sounding recital that ignores the follow-up.

The Speaking band is reported from 1.0 to 6.0, aligned to CEFR. Raters listen for a few broad qualities: whether they can understand you easily, whether your speech flows without long unnatural pauses, whether your answer holds together as one connected idea, and whether your grammar and vocabulary support what you are trying to say.

Accent alone does not lower your band. Unclear pronunciation that stops the rater from understanding you does. Ambition does not raise your band either — an ambitious sentence you cannot finish costs more than a simple sentence you can.

Short, frequent drills beat one long weekly session. Ten minutes of speaking every day trains fluency and reduces the fear of the microphone in a way that a two-hour block on Sunday cannot. Recording yourself and listening back is the fastest single improvement for most candidates — it forces you to hear the pauses and repetitions you did not notice while speaking.

For the interview task specifically, practise answering follow-up questions. Ask a partner or a study group to react to your answer with a real follow-up. Do the same task twice: once, then again with a different follow-up. You will feel the difference between reciting and thinking on your feet.

You can record answers on our platform and get AI feedback on intelligibility, fluency, coherence and language use. Feedback points at specific moments in your recording, not just a general grade, so you know exactly where a stumble or a grammar slip cost you.

Here is the honest note: until you record real answers, your Speaking level is simply not yet assessed. We never invent a Speaking band from your reading score or from how much you have typed. If you skip the recording tasks, your Speaking line reads not yet assessed. That is the truth, and it is more useful than a made-up number.