1.1 TOEFL iBT Structure, Timing & 2026 Adaptive Design

TOEFL iBT Structure, Timing, and Adaptive Design

The current TOEFL iBT measures Reading, Listening, Writing, and Speaking in a fixed sequence. The format described here is the version used after January 21, 2026. That date matters: many older books and web pages still describe two 700-word reading passages, multi-minute lectures, an Integrated Writing essay, and four legacy Speaking tasks. Those are not the operational task families in the current test.

ETS describes the updated test as taking about two hours when directions, equipment checks, and other administration are included. Its content table lists a shorter base testing time. Candidates should therefore treat the published section figures as planning anchors, not as a promise that every administration will have an identical minute or item total.

Official section order and base table

The 120 base items are not 120 conventional multiple-choice questions. The Writing total includes ten machine-scored Build a Sentence items and two constructed responses. The Speaking total includes seven repeated sentences and four interview answers. ETS also notes that directions are outside the base time and that time and items may vary as the test adapts.

What “two-stage adaptive” means

Reading and Listening use multistage adaptation. Each begins with a first-stage router. Performance in that first part is then used to select content for the second part so that task difficulty better matches the test taker’s demonstrated ability. ETS’s 2026 blueprint identifies a lower or upper second module and gives combined timing ranges for the router and selected module.

That official description supports three practical conclusions:

1. Work accurately from the start; the first stage contributes evidence used to choose the second-stage route.
2. Do not panic if the second stage feels easier or harder. Perceived difficulty is not a reliable score calculator.
3. Expect the section to present different mixes of difficulty across test forms.

Do not build strategy around unsupported claims. ETS does not tell candidates that there are three second-stage branches, publish a score ceiling for a route, disclose operational cut scores, or say that one hard item is worth a fixed multiple of an easy item. The score is a scaled proficiency estimate, not a simple percentage, but the safest candidate behavior is still ordinary: answer every item as carefully as the clock permits.

Current task distribution

The official blueprint provides useful task-level allocation ranges. Reading contains 30 Complete the Words items, with the remaining 20 divided between Read in Daily Life and Read an Academic Passage sets. Listening contains 15–19 Listen and Choose a Response items, 10 conversation questions, 6–10 announcement questions, and 8–16 academic-talk questions. Writing contains ten sentence reconstructions plus one email and one discussion response. Speaking contains seven repetitions plus four interview answers.

These allocations explain where preparation time belongs. Complete the Words is 60% of the Reading base count, while brief response selection is the largest single Listening family. That does not make extended comprehension unimportant: academic passage and talk sets test inference, organization, vocabulary, and relationships among ideas at higher proficiency levels.

Continuous testing and completion rules

ETS states that the current TOEFL iBT has no scheduled break. Approved accommodations can change timing or break conditions, but ordinary administrations run continuously. The current test order also places Writing before Speaking, so old stamina plans that expect Speaking before Writing are out of date.

To receive scores, a test taker must answer at least one Reading question, one Listening question, one Speaking task, and one Writing task. That is only a scoring-eligibility minimum, not a strategy: unanswered items cannot demonstrate proficiency. Plan to complete every section.

A sound adaptive-test mindset

Use the visible directions and timer as the authority on test day. Move through each item with enough time to read or listen for meaning, verify the response, and continue. Do not try to diagnose the route while testing; doing so consumes attention without providing actionable information. The updated design rewards the same language behaviors it claims to measure: lexical and grammatical processing, accurate comprehension, context-sensitive communication, and clear spoken and written production.