TOEFL Speaking Templates (2026): Tasks 1–4, Timing Drills & Scoring Criteria

The four tasks at a glance

The whole section runs roughly 16 minutes and is scored 0–30, scaled up from a raw 0–4 on each task. Because there are only four tasks, one badly mishandled response moves the section score noticeably — which is an argument for a reliable structure on every task rather than brilliance on one.

Note the pattern: only Task 1 asks what you think. Tasks 2, 3 and 4 ask you to report what other people said. Volunteering your own opinion in the integrated tasks is one of the most common ways students spend their limited seconds on content that earns nothing.

How Speaking is actually scored

Responses are evaluated using a combination of human raters and automated scoring. The practical consequence is that consistency pays: a clear, complete, well-organised answer at a moderate pace scores better than an ambitious one delivered in a rush with two abandoned sentences in it.

Topic Development is where templates earn their keep, and also where they can hurt. Covering every required element is exactly what a template guarantees; padding with memorised phrases that add no content is exactly what the criterion penalises. Keep your fixed language to transitions — "According to the reading", "The professor illustrates this with" — and let everything else come from the prompt in front of you.

If you are applying for teaching assistantships, note that some programmes set a separate minimum on the Speaking section rather than only on the total. For those applicants this section quietly matters more than any other, so check the requirement before deciding where to spend prep time.

Task 1 template — the independent response (45 seconds)

Two developed reasons beat three thin ones. With 45 seconds, three reasons leaves about twelve seconds each, which is not enough to develop anything, and Topic Development rewards development. If you can only think of one genuine reason, split it into a cause and its consequence rather than inventing a weak second one.

Your examples do not need to be true. A plausible, specific illustration does the job, and specificity is what makes a response sound developed — "when I was preparing for my semester exams in a noisy hostel" earns more than "sometimes in some situations". In your 15 seconds of prep, do not write sentences; write your position and two keywords. Sentences you cannot finish reading are worse than keywords you can speak from.

• 0–5 sec — State your position in one direct sentence, reusing the prompt's wording: "I would prefer to <option A>, mainly for two reasons."
• 5–22 sec — First reason plus a specific detail: "First, <reason>. For example, when I <a concrete personal example>, I found that <what happened>."
• 22–40 sec — Second reason plus a detail: "Second, <reason>. In my case, <specific detail>."
• 40–45 sec — One short closing clause if time remains: "So overall, <option A> works better for me." Drop it without hesitation if you are still finishing reason two — an unfinished second reason costs more than a missing conclusion.

Task 2 template — campus reading plus conversation (60 seconds)

The reading passage is scenery. It exists so your summary has context, and it deserves about ten seconds — students who spend twenty-five seconds retelling the announcement routinely run out of time before the speaker's second reason, which is the content that actually gets marked.

Listen specifically for the two reasons, and for the detail attached to each. The conversation is designed so one speaker holds a clear opinion; identify that speaker in the first few seconds of the audio and note only their side. Their reasons are almost always signposted with contrast markers — "but", "the thing is", "besides" — and those markers are your cue to write.

• Note-taking grid: split your page in two. Left — the proposal and the two reasons the announcement gives. Right — which speaker has the strong opinion, whether they agree or disagree, and their two reasons.
• 0–10 sec — The proposal and its stated reason: "The university plans to <change>, because <reason from the reading>."
• 10–16 sec — The speaker's stance: "The <man/woman> disagrees with this plan for two reasons."
• 16–36 sec — Their first reason with the detail they gave: "First, he points out that <reason>. He explains that <specific detail from the conversation>."
• 36–58 sec — Their second reason with detail: "Second, he mentions that <reason>, because <detail>."
• Never add your own view. The task asks what the speaker thinks, and your opinion is off-task content occupying seconds you need.

Task 3 template — academic concept plus lecture (60 seconds)

The lecture carries the weight here, not the reading. Students who define the term for thirty seconds and then compress the professor's example into fifteen have inverted the task — the definition is the setup, and the example is the answer. Keep the definition to roughly one sentence even if you understood it perfectly.

Use the professor's specifics: names, numbers, and the sequence of events. If the lecture describes an experiment with birds, say birds. Generic paraphrase ("the professor gives an example about animals") reads as not having understood the lecture, and it is thinner content in a section that scores development.

• Note-taking grid: left — the term and a compressed definition, in your own shorthand. Right — the professor's example, in sequence: what happened first, what happened next, what it showed.
• 0–15 sec — Define the term from the reading in your own words: "The reading explains <term>, which refers to <definition>."
• 15–20 sec — Bridge to the lecture: "In the lecture, the professor illustrates this with an example about <topic>."
• 20–40 sec — The first half of the example, in the order it was given: "He describes how <what happened>. As a result, <outcome>."
• 40–58 sec — The rest of the example and the connection back to the term: "Then <what followed>, which shows how <term> works in practice."

Task 4 template — lecture only (60 seconds)

Task 4 has the shortest preparation time — 20 seconds — and no reading passage to anchor you, which makes note quality decisive. Write the two point-headings first, in two or three words each, and only then fill in examples. Full-sentence notes are a trap here: you cannot write them fast enough, and you cannot read them while also speaking naturally.

The listening itself is the skill under test more than the speaking is. If Task 4 is consistently your weakest, the fix is usually listening practice with note-taking — academic lectures, played once, summarised aloud from notes — rather than more speaking templates.

• Note-taking grid: main topic across the top, then two columns for the two points, examples underneath each. Almost every Task 4 lecture is structured as one concept with two aspects, types, effects or strategies — expect that shape and your notes will fit it.
• 0–8 sec — Name the topic and the structure: "The lecture discusses <topic> and describes two <types / strategies / effects>."
• 8–32 sec — First point with its example: "The first is <point>. The professor explains that <detail>, and gives the example of <example>."
• 32–58 sec — Second point with its example: "The second is <point>. In this case, <detail and example>."
• If you missed one of the two points, expand the one you have with its example rather than going silent or admitting the gap. Content you can deliver scores; content you cannot does not.

Timing drills that fix the real problems

• Record everything and listen back. This is the single most skipped and most valuable habit in Speaking prep — students hear their own long pauses, flat intonation and abandoned sentences immediately, and almost never notice them while speaking.
• A workable two-week routine: days 1–3 learn the four templates until the transitions are automatic; days 4–9 two tasks a day, timed and recorded, alternating types; days 10–12 full Speaking sections back-to-back, since fatigue across four tasks is its own skill; days 13–14 re-record your three weakest responses.
• Practise out loud, standing, at speaking volume. Rehearsing silently in your head builds none of the muscle the test measures, and it is how students arrive at the exam having "prepared" without ever having spoken for 60 seconds continuously.
• Practising against something that asks follow-up questions — a study partner, a teacher, or an AI interviewer — helps more than solo repetition once your templates are solid, because the skill under pressure is producing structured speech at will, not reciting a rehearsed answer.

What caps students who already speak good English

• Memorised content. Rehearsed sentences that do not answer the prompt are penalised, and raters recognise them easily. Keep memorisation to transitions and structure.
• Giving your own opinion on Tasks 2, 3 and 4. These ask what the speaker or professor said. Your view is off-task, and it eats seconds you need.
• Over-summarising the reading in Tasks 2 and 3. The reading is context; the audio is the content that gets marked.
• Trailing off at the end. A response cut mid-word signals poor planning. Watch the clock at 40 seconds and shorten your last point to land it cleanly.
• Long opening pauses. Four seconds of silence at the start is four seconds of a 45-second answer, and it sets a hesitant tone for everything after it.
• Speed. Nervous candidates accelerate, and clarity — which Delivery actually scores — falls with it. Slightly slower than feels natural is usually right.
• Ignoring the microphone. On the home edition especially, test your mic and speak at a normal conversational volume; a recording the rater struggles to hear cannot earn marks for content it does not capture.
• Practising only Task 1 because it is comfortable. Three of the four tasks are integrated, and they are where most students actually lose the section.

Frequently asked questions