Preparing with a pre-2026 TOEFL video, book, or template? Stop before you memorize the wrong exam. For TOEFL iBT tests taken on or after January 21, 2026, the Speaking section uses Listen and Repeat and Take an Interview—not the former independent and integrated Speaking tasks.

Direct answer: the current Speaking task types

Official ETS fact: The current Speaking section has two task types. ETS’s Speaking section page describes Listen and Repeat as exact repetition of short spoken sentences and Take an Interview as a simulated academic or campus interview about experiences and opinions.

Current TOEFL Speaking structure for tests on or after January 21, 2026
Task type Items in the current technical blueprint What you do What the task broadly samples
Listen and Repeat 7 Hear a short sentence and repeat it as accurately and intelligibly as possible. Rapid processing of spoken English, retention of linguistic form, repetition accuracy, fluency, and intelligibility.
Take an Interview 4 Answer a simulated interviewer’s questions about an academic or campus-related situation. Immediate, meaningful speech; clear development; intelligibility; pace; vocabulary; grammar; and organization.
Speaking total 11 Complete both task types in one recorded Speaking section. ETS lists an approximate base time of eight minutes, excluding directions.

The 7-plus-4 split comes from the current TOEFL iBT® Technical Manual and 2026 blueprint and specifications document. The simpler ETS test-content overview gives the public-facing total: 11 items and an approximate base time of eight minutes.

The practical consequence: do not spend your main preparation time learning the retired four-task system. Build two separate abilities instead: accurate one-hearing repetition and responsive interview speaking.

What Listen and Repeat tests

Official ETS fact: You hear short sentences and repeat them exactly. ETS describes the target as processing spoken English and producing speech that is accurate and clearly intelligible. Its technical manual adds that items appear in a scenario and become longer and more syntactically complex as the sequence continues.

This task exposes two problems that learners often blur together:

  • Repetition accuracy: Did you preserve the words, word order, grammatical endings, and key function words?
  • Intelligibility: Could another listener understand what you actually said without reading the sentence?

A paraphrase can be perfectly understandable and still be an inaccurate repetition. Conversely, you may remember every word but pronounce several words so unclearly that the response is hard to follow. Diagnose those failures separately.

Original Listen and Repeat practice item

Original practice—not an ETS question

Situation: A staff member is giving a visiting student information about a campus building.

The student laboratory closes early on Friday afternoons.

One-take practice: Have a partner or text-to-speech tool play the sentence once while you look away from the text. Repeat it once into a recorder. Then reveal the sentence and compare your recording.

  • Did I keep every content word?
  • Did I keep small words and endings such as the, on, and plural or tense markers?
  • Did I preserve the original word order rather than repair the sentence from memory?
  • Is the recording understandable at normal volume without the transcript?

That is enough for a first diagnosis. For a deeper progression covering sentence length, memory load, pronunciation, and error logging, continue to TOEFL Listen and Repeat: Accuracy, Memory, and Pronunciation.

What Take an Interview tests

Official ETS fact: Take an Interview is a simulated interview connected to academic or campus situations. ETS says you answer questions about your experiences and opinions. The technical manual describes a progression from relatively factual or personal responses toward broader questions that may require an opinion, explanation, prediction, or narration.

The task is not just “speak for as long as possible.” It checks whether you can understand the new question, begin without a long collapse into silence, answer the actual angle, and develop a coherent response at a natural pace. More words do not rescue an off-topic answer.

Original Take an Interview practice sequence

Original campus practice—not an ETS question or official mock

Situation: A university library is reviewing how students use its study spaces.

  1. What kind of study space do you use most often on campus?
  2. What feature of that space helps you study effectively? Give one example.
  3. The library can improve either quiet rooms or group-work rooms this year. Which should it prioritize, and why?
  4. Some students want every study room to require an online reservation. Do you agree? How could that policy affect students with different schedules?

One-take practice: Record one answer per question. Do not rewrite an earlier answer when the angle changes. Question 2 asks for evidence, Question 3 asks for a choice and justification, and Question 4 asks you to evaluate a policy and its consequences.

“Adapting to follow-up questions” here means tracking the changed demand across successive questions. The official sources reviewed for this package do not say that the interview branches dynamically according to your individual answer, so do not build a strategy around guessing a hidden branching algorithm.

For fuller interview-specific work on response openings, development, and question changes, use TOEFL Take an Interview: Questions and Practice.

How the section is delivered

Official ETS facts:

  • The TOEFL iBT is offered on a computer at an authorized test center and as a Home Edition monitored by a live human proctor.
  • In Speaking, you speak into a microphone and your responses are recorded.
  • The scenarios are academic or campus-related, and ETS says specialized background knowledge is not required.
  • You may hear native-speaker accents from North America, the United Kingdom, New Zealand, or Australia.
  • The public overview lists 11 Speaking items and an approximate base time of eight minutes. Directions are not included in that base-time figure.

The current public Speaking page, blueprint, and technical manual reviewed for this hub do not establish the former per-task pattern of 15 or 30 seconds to prepare and 45 or 60 seconds to speak. Those numbers belong to legacy task advice unless ETS explicitly attaches them to a current task in a current resource. Practice quick comprehension and one-take delivery without pretending an old timer is still official.

For test-day familiarization, use ETS’s current full-length sample test page rather than relying on screenshots from an older preparation book.

Build a diagnostic practice plan

Editorial guidance—not an official score model: Start with one recorded baseline before choosing drills. Use several original repetition sentences that rise in length, followed by one four-question interview sequence. Review the recording once for content and once for delivery. Then name one primary bottleneck and one secondary bottleneck.

How to identify the main TOEFL Speaking practice problem
Possible main problem What it sounds like Simple diagnostic Best next focus
Accurate repetition You omit or replace words, change word order, or lose small grammatical words even when the recording is understandable. Compare your recording with the written sentence word by word. Mark omissions, substitutions, additions, and order changes. Use shorter-to-longer one-hearing repetition sets and an error log; route deeper work to TF02.
Intelligibility You remember the sentence or idea, but endings disappear, words merge unpredictably, or a listener cannot identify key words. Ask a listener to write what they hear without seeing the prompt, or transcribe your own recording after a delay. Work on the few sounds, stress patterns, endings, or pacing habits that repeatedly hide meaning—not on changing your accent wholesale.
Immediate response You understand the interview question but repeatedly stall, restart, or search for a perfect opening before giving any answer. Play five varied questions. After each one, begin with a direct position or fact before planning the rest. Use the response-start drills in Answer Questions Faster in English and interview work in TF03.
Answer development You answer the question directly but stop before explaining, supporting, or illustrating the point—or you add unrelated detail. Label each sentence in a response as answer, reason, example, consequence, or tangent. A response made mostly of tangents needs structure, not more time. Practise one controlled elaboration move at a time with Give Longer Answers Without Rambling.
Adapting to follow-up questions You recycle the same prepared answer after the interviewer changes from experience to reason, choice, prediction, or consequence. Before each response, state the new job in a few words: “give an example,” “choose,” “predict,” or “evaluate.” Then answer that job. Use varied sequences from Conversation Questions, then apply the current interview structure in TF03.
  • Record one take; do not pause and rebuild the answer.
  • Separate content errors from delivery errors.
  • Choose one main problem for the next three sessions.
  • Keep evidence: the prompt, recording, transcript or notes, and one-line diagnosis.
  • Do not convert this checklist into a predicted TOEFL score.

English pronunciation

If your words are accurate but difficult to understand, use the English pronunciation hub to identify the relevant component—such as word stress, sentence stress, rhythm, endings, or a sound contrast. Intelligibility is the goal; an imported “native accent” is not required by the current task description.

A two-week schedule

Editorial guidance: The schedule below assumes roughly 25–35 focused minutes a day. That duration is a study default, not an ETS requirement. Keep every recording, because visible error patterns are more useful than a vague feeling that you “need more fluency.”

Fourteen-day current-format TOEFL Speaking practice schedule
Day Main work Evidence to save Route or decision
1 Record a baseline: use the original sentence in this hub plus six similar campus statements you create, then complete the four-question interview sequence. Audio, repetition error counts by type, and one note for each interview answer. Choose the single biggest bottleneck from the diagnostic table.
2 Practise short repetition items. Prioritize exact words, order, and grammatical endings before speed. A marked transcript showing omissions, substitutions, additions, and order changes. Use the progression in TF02.
3 Increase sentence length gradually. Repeat once after one hearing; do not repair from the transcript first. The longest level at which accuracy remains stable, plus the first recurring failure. Stay below overload long enough to remove the recurring error, then increase length.
4 Re-record selected repetition items for intelligibility: clear key words, endings, stress, and steady phrasing. A blind listener transcript or your own delayed transcription. Route repeated clarity problems to pronunciation work; keep memory work separate.
5 Answer ten mixed campus questions with a direct first sentence. Avoid building a perfect introduction. Mark stalls, restarts, and answers that begin without addressing the question. Use Answer Questions Faster in English for response-start practice.
6 Take five short answers and add one relevant reason, example, or consequence to each. Sentence labels: answer, reason, example, consequence, or tangent. Use Give Longer Answers Without Rambling.
7 Practise two four-question interview sequences whose angles change from fact to support, choice, and evaluation. A one-line “question job” before each answer and a note on any recycled response. Use TF03 for deeper sequence practice.
8 Run one mixed 7-plus-4 practice set without stopping between items except for the practice interface’s prompts. One full recording and only three notes: strongest behavior, biggest loss, next fix. Do not chase every small error at once.
9 Spend the session on the primary bottleneck found on Day 8. Before-and-after examples using comparable prompts. Choose TF02 for repetition or TF03 for interview performance.
10 Train the secondary bottleneck while preserving yesterday’s improvement. One repetition set and one interview answer demonstrating both targets together. Reduce task difficulty if combining the targets causes both to collapse.
11 Practise microphone discipline: stable distance, audible volume, one take, and no self-interruption after a minor mistake. Recordings made with the equipment and room setup you plan to use for practice. This is delivery rehearsal, not an official equipment certification.
12 Use the official current full-length sample test to check the live interface and current content structure. Notes on any format surprise and the same five diagnostic categories used earlier. Prefer the current ETS experience over an undated third-party simulation.
13 Repeat a mixed practice set. Apply only the two changes that produced the clearest improvement. A Day 1 versus Day 13 comparison using specific observable differences. Keep successful routines; discard rituals that did not change performance.
14 Review the recordings, choose the next two-week priority, and check how the current reporting scale works. A one-page plan naming one task target, one delivery target, and the evidence you will collect. Use TOEFL Speaking Scores and Strategy for the 1–6 Scale for score interpretation—not a homemade conversion.

Current scoring path

Official ETS facts: Your Speaking responses are recorded rather than scored by an interviewer in the room. ETS’s technical manual describes proprietary AI scoring used with human scoring to improve accuracy and consistency. It also describes certified human-rater quality-control review and human intervention when automated scoring has low confidence or a response is otherwise difficult to score.

  1. You produce the responses: seven repetition items and four interview answers in the current technical design.
  2. ETS evaluates the responses: automated models and human scoring processes are used; this hub does not invent a percentage weight for either.
  3. You receive a Speaking section score: for tests from January 21, 2026, ETS reports section scores on a 1–6 scale in half-point increments.
  4. The transition also affects the overall score: ETS says score reports include a comparable overall 0–120 score during a two-year transition period. That is not permission to manufacture a FunFluen-to-TOEFL conversion.

ETS does not set one universal passing score; institutions decide their own requirements and may update admissions pages at different times. Check the institution’s current requirement and your official report. For the score scale, interpretation, and strategy boundary, use TOEFL Speaking Scores and Strategy for the 1–6 Scale.

Which older TOEFL advice is obsolete

The January 2026 change did not merely rename the old Speaking tasks. It replaced the Speaking task set. Keep transferable language skills, but remove retired task assumptions.

Migration from legacy TOEFL Speaking advice to the current format
Retired or legacy advice Why it is no longer current-format guidance Use this replacement
Prepare four Speaking tasks: one independent and three integrated. That structure applies to tests before January 21, 2026, not the current Speaking section. Prepare seven Listen and Repeat items and four Take an Interview questions as the current technical blueprint.
Memorize universal 15- or 30-second preparation routines and 45- or 60-second response templates. Those timing patterns belong to former task instructions. The current sources reviewed for this hub do not attach that legacy pattern to the two new Speaking tasks. Practise rapid comprehension, direct starts, accurate repetition, and one-take responses; use the current ETS interface for exact test-day cues.
Build Speaking preparation around reading-and-listening integrated summaries. The current Speaking task list is Listen and Repeat plus Take an Interview. Move repetition accuracy and interview response development to the center of the plan.
Treat the independent opinion task as the only personal-response format. Take an Interview now asks about experiences and opinions across a sequence of questions. Practise direct personal answers, support, choices, explanations, predictions, and consequences.
Use the former 0–30 Speaking scale as the only current reporting language. For tests from January 21, 2026, section scores are reported on the 1–6 scale in half-point increments. Read the current score report and the receiving institution’s requirement; do not perform an unofficial conversion.
Read the existing FunFluen TOEFL iBT fluency-drills page as a current format map. That page is legacy content. General fluency exercises may transfer, but retired task names, timing, templates, or sequencing must not be presented as current. Use this hub for current format, TF02 for Listen and Repeat, TF03 for Take an Interview, and TF04 for scores.

Transferable skills still matter: clear pronunciation, stable pacing, direct answers, relevant support, and the ability to continue after a small mistake. The obsolete part is the old exam wrapper—not every speaking skill practised inside it.

Official resources

Use current ETS pages as the format authority. Third-party material can add practice, but it should not override a dated official task list.

Before following a TOEFL Speaking lesson, check two things: does it explicitly apply to tests on or after January 21, 2026, and does its task list match Listen and Repeat plus Take an Interview? If either answer is no, treat it as legacy until proven current.