TOEFL Speaking Templates: What to Actually Say in Each Task

A TOEFL speaking template only helps if it matches the task you actually get in 2026, and most templates online do not. The TOEFL iBT Speaking section has two task types now: Listen and Repeat and Take an Interview. The old four-task structure, with its separate independent and integrated questions, was retired. If a guide still talks about Task 1 through Task 4, it is describing an exam that no longer exists, and applying its templates to the current test will not help your score.

That matters because the two current tasks reward completely different skills, and one of them cannot be templated at all. Take an Interview rewards how you organize an opinion, so a flexible answer skeleton helps there. Listen and Repeat rewards how accurately you reproduce a sentence you just heard, so no skeleton applies; what helps instead is a specific way of holding the sentence in memory. Mixing the two up is the most common mistake in speaking prep, and it wastes the seconds you actually have to answer (ETS).

Key Takeaways

• The 2026 Speaking section has two task types: Listen and Repeat (7 questions, up to 35 points) and Take an Interview (4 questions, up to 20 points), for 55 points total.
• Take an Interview is scored on Fluency, Intelligibility, Language Use, and Organization. A 4-part answer skeleton directly targets the Organization and Language Use dimensions.
• Listen and Repeat is scored on Fluency, Intelligibility, and Repeat Accuracy. Repeat Accuracy rewards exact reproduction, not paraphrase, so no template applies; a chunking and stress-mapping strategy does.
• Your overall Speaking band is the sum of all 11 item scores out of 55, converted to a 1-6 band.

The Two 2026 Speaking Tasks, and Why They Need Different Preparation

The table below is drawn from the current scoring specification, not from a general description of the test.

Each item is scored from 0 to 5 on its dimensions, and the item score is the average of those dimensions. All 11 item scores are added together for a total out of 55. That total is divided by 55, and the result is converted to the 1-6 band scale used for the whole exam. A single strong Take an Interview answer does not average out several weak Listen and Repeat items, because both task types feed the same total; strengthening whichever one is currently weaker moves the overall Speaking band more than polishing an already-strong task.

Listen and Repeat carries more than half the available points, which is a reason to take its preparation seriously even though it is not a task you can template.

Take an Interview: An Adaptable 4-Part Skeleton

Take an Interview asks you to respond to a short series of spoken questions about a familiar academic or campus situation, with your own opinion or experience. There is no preparation time before each question; you answer immediately, with roughly 45 seconds per response, and the questions get progressively more demanding as the set continues (MySpeakingScore). Because the question changes every time and there is no moment to plan, memorizing full answers does not work, and the AI scoring system is built to notice generic, disconnected responses. What transfers between questions is the shape of a well-organized answer, not its exact words, and that shape has to be automatic enough to use while you are still thinking of the content.

The skeleton below has four parts. Treat it as a structure to adapt, not a script to recite; the words in brackets are where you insert content specific to the question you are asked.

1. Direct answer. State your position in one sentence, using the words of the question where natural. Example pattern: "I would prefer [option], because [short reason]." This sentence alone should make your position unambiguous to a listener who hears nothing else.
2. Reason. Expand the short reason from step one into a full sentence or two. Example pattern: "The main reason is that [specific factor], which matters to me because [personal consequence]."
3. Specific example. Give one concrete detail: a time, a place, a person, or a number. Example pattern: "For example, when [specific situation], [what happened]." A specific example is what separates a Language Use score that rewards precise vocabulary from a generic one that does not.
4. Closing. Return to your position in different words, or add a brief qualification. Example pattern: "That is why, for me, [restated position] makes more sense than the alternative."

How the Skeleton Maps to the Rubric

Organization is scored on whether a listener can follow your answer without effort. A direct answer stated first, followed by a reason and an example, gives the response a shape a listener can track from the first sentence. Language Use is scored on range and accuracy of vocabulary and grammar, not on using difficult words; a specific example naturally pulls in more precise nouns and verbs than a general statement does, which is where Language Use scores tend to improve. Fluency and Intelligibility are shared with Listen and Repeat and depend on your speech itself, not on the skeleton, so the structure above supports two of the four dimensions directly and does not interfere with the other two.

Transition Language Between Parts

Natural connections between the four parts matter more than the content of any single part, because abrupt jumps between ideas are what a listener notices first. A short bank of transition phrases, used naturally rather than mechanically, is enough to link the parts:

• Into the reason: "This is mainly because...", "The main factor for me is..."
• Into the example: "For instance...", "A specific case was when..."
• Into the closing: "Overall...", "So in the end..."

Reusing the same two or three transitions across your practice answers is fine. Reusing the same full sentences for every topic is not, because the AI scoring system compares your answer against the specific question you were asked, and a disconnected answer scores lower on Organization even if the grammar inside it is correct.

Practicing Without Preparation Time

Since there is no thinking time before you must start speaking, the skeleton only helps if it is practiced until choosing a position and starting the first sentence happens without conscious planning. Two habits build that speed. First, practice starting your answer within two or three seconds of hearing a question, even with a weak or uncertain position, because a direct answer that arrives late costs more on Organization than an imperfect one that arrives on time. Second, treat the first sentence as the part you say automatically, and let the reason and example develop while you are already speaking; by the time you finish your direct answer, you have used those few seconds of speech to decide what comes next, which is the only preparation time this task actually gives you.

With roughly 45 seconds per question, the four parts do not need to be long. A direct answer of one sentence, a reason of one or two sentences, one specific example, and a short closing comfortably fit the time, and rushing to add more content usually hurts Fluency more than it helps Language Use.

If you want to check whether your Take an Interview answers actually hit these dimensions, practice with AI scoring on real prompts and read the feedback against each dimension separately, not just the overall number.

Listen and Repeat: Why a Template Cannot Help, and What Does

Listen and Repeat plays a sentence once, with no text shown on screen, and you repeat it as closely as possible after a short signal. Each of the seven sentences gets slightly more response time as the set progresses, moving from about 8 seconds for the first two sentences to about 12 seconds for the last two, which reflects the sentences themselves getting longer, not extra thinking time (MySpeakingScore). Repeat Accuracy, one of the three scored dimensions, measures how closely your output matches the original wording. Rephrasing the sentence, even correctly, lowers this score, because the task is not testing whether you understood the idea; it is testing whether you can reproduce spoken English accurately after hearing it once. This is the opposite of Take an Interview, where using your own words is expected.

Because every sentence is different, there is no answer skeleton to reuse here. What transfers between items instead is a way of processing what you hear, and three habits make a measurable difference.

Chunk the Sentence by Meaning, Not by Word

Trying to hold an entire sentence in memory word by word overloads short-term memory, especially for longer sentences. Breaking the sentence into two or three meaning-based chunks as you hear it, rather than as a list of individual words, is easier to retain and easier to reproduce in order. A sentence like "The professor announced that the deadline for the assignment had been moved to next Friday" naturally splits into three chunks: "the professor announced," "that the deadline for the assignment," "had been moved to next Friday." Practicing this kind of chunking on unfamiliar sentences, not only on ones you already know, builds the skill the task actually requires.

Match Stress and Intonation, Not Just Words

Fluency and Intelligibility, the two dimensions Listen and Repeat shares with Take an Interview, are affected by rhythm as much as by word choice. A repeated sentence with all the correct words but flat, even stress on every syllable sounds less natural than one with the original sentence's stress pattern, and it scores lower on Intelligibility. As you listen, notice which one or two words in the sentence carry the main stress, usually the words that carry the most meaning, and reproduce that emphasis rather than speaking every word at the same volume and pace.

Use the First Instant After the Signal, Not a Planning Pause

Because you respond right after a short signal with no extra thinking time built in, the chunking has to happen while you are still listening, not afterward. Practicing with your eyes closed on unfamiliar sentences, actively grouping the words into chunks as they arrive rather than waiting until the sentence ends, trains this in real time. This is a listening habit, not a content strategy, and it is the closest thing to a repeatable technique this task allows, because it works the same way regardless of what the sentence says.

For a broader walkthrough of how the whole 2026 exam is structured and scored, see our complete guide to the TOEFL iBT 2026; this article intentionally goes deeper into these two tasks rather than repeating that overview.

A Combined Practice Routine

Because the two tasks need different preparation, alternating between them in a single practice session, rather than drilling one for a full session, tends to produce steadier gains across both. A structure that works for most learners:

1. Warm up with two Listen and Repeat items, focusing only on chunking, without worrying about the score.
2. Answer two Take an Interview questions using the 4-part skeleton, starting your first sentence within two or three seconds each time.
3. Return to three more Listen and Repeat items, this time focusing on stress and intonation.
4. Answer two more Take an Interview questions, this time without looking at the skeleton written down, to check whether the structure has become automatic.
5. Review the AI feedback for both task types separately, since a low Organization score and a low Repeat Accuracy score point to different fixes.

Ten to fifteen minutes of this alternating pattern, repeated several times a week, builds both skills without letting one crowd out the other, which is a common problem when learners over-practice the task that already feels comfortable.

Frequently Asked Questions

Is there a template for Listen and Repeat?

No. Repeat Accuracy specifically rewards reproducing the sentence you heard, not restating it in your own structure, so a fixed template would actively work against the score. The chunking and stress-matching habits described above are the closest equivalent, but they are memory and pronunciation techniques, not sentence templates.

Can I use the same opening sentence for every Take an Interview question?

The shape of the opening sentence (state your position, using the question's own wording where it fits) can repeat, but the content inside it must change with every question. An opening sentence with no specific content, reused word for word, is exactly what the Organization and Language Use dimensions are built to catch as generic.

How many points is the Speaking section worth toward my overall score?

Speaking is one of four sections, each converted to its own 1-6 band, and the overall score is the average of all four section bands. Within Speaking itself, Listen and Repeat contributes up to 35 of the 55 total points and Take an Interview up to 20, so both tasks matter, but Listen and Repeat has more raw weight in the section total.

What is the fastest way to know which task is weaker for me?

Score-level feedback alone will not tell you, because a single Speaking band mixes both tasks together. Item-level feedback that separates Fluency, Intelligibility, Repeat Accuracy, Language Use, and Organization is what shows which specific task and which specific dimension needs work.

Put This Into Practice

A template only moves your score if you practice it against real questions and get feedback on whether it worked. Take a free practice session covering both Speaking tasks, and check your item-level results against the dimensions in this article: Organization and Language Use for Take an Interview, Repeat Accuracy for Listen and Repeat, and Fluency and Intelligibility for both.