TOEFL iBT 2026 timing: how the clock works in each section

Reading: one clock per module

Reading is timed by module. Each of its two modules has a single countdown for all of its tasks, and you decide how to divide that time between them.

A common mistake is to spend too long on Complete the Words at the start and then rush the academic passage, where each question takes more reading to answer.

Listening: a limit on every question

Listening has no section clock for you to manage. Each recording plays once, and then every question has a short time limit of its own: about 20 seconds for most questions and about 30 seconds for questions on an academic talk.

When the time for a question ends, the test moves on, so there is no way to save time on easy questions and spend it on hard ones. Decide, answer and move forward.

Writing: three tasks, three clocks

The writing tasks are timed separately, and time left over from one does not carry into the next. The Build a Sentence set has about 7 minutes in total, Write an Email has 7 minutes and Write for an Academic Discussion has 10 minutes.

Seven minutes is short for an email that has to cover several points. Reading the task carefully, writing directly and leaving a minute to check it is a realistic plan.

Speaking: timed responses, no preparation time

Speaking is timed response by response. In Listen and Repeat you have about 8 to 12 seconds to say each sentence, depending on its length. In Take an Interview you have 45 seconds for each answer.

Recording starts and stops automatically and there is no preparation time, so the skill being tested is answering in the moment, not delivering a prepared speech.

Where these figures come from

ETS publishes time estimates for each section rather than a time for every item. The per-question listening times and the Build a Sentence time given here come from test-preparation reporting, which is why they are marked "about".

Practise on the clock

Untimed practice hides exactly the problems these clocks create. Work with a timer from the start, and sit at least one full test in a single sitting so that moving from one kind of clock to the next feels ordinary.