TOEFL 2026 note
Sixty per cent of Reading is the task you think is a warm-up
Complete the Words is 30 of the 50 items in TOEFL 2026 Reading. Most people spend their Reading practice on academic passages, which are a fifth of the section.
The weighting almost everyone has backwards
| Task | Items |
|---|---|
| Complete the Words | 30 |
| Read in Daily Life | 5–15 |
| Read an Academic Passage | 5–15 |
| Total | 50 · 30 min |
Complete the Words is thirty of the fifty items. The two tasks that look like “real” reading — the practical documents and the academic passage — split the remaining twenty between them.
Most people preparing for this test have the weighting exactly backwards. They spend their Reading practice on academic passages, because that is what TOEFL Reading has meant for twenty years, and treat the word-completion items as a gentle opening. It is not an opening. It is the section.
What the task actually tests
You get a passage with gaps, and each gap shows you the first letters of the missing word. You complete it. The CEFR range ETS assigns it is B1–C1+, which is worth noticing — this is not an elementary task.
What it tests is not vocabulary in the flashcard sense. You are given the beginning of the word, so recognition is not the problem. The problem is whether you can hold enough of the sentence's grammar and meaning in place to know which form of a half-recognised word belongs in the gap — the noun or the verb, singular or plural, the collocation that actually goes with the preposition sitting three words later. That is a different skill from reading comprehension, and it does not improve much as a side effect of reading more.
Why it is the highest-leverage task in the section
It is 60% of the items, so the same improvement moves your band further here than anywhere else in Reading. It is fast to practise — thirty gaps in the time one academic passage takes. And its feedback is clean: wrong part of speech, wrong number, right idea and wrong collocation are all visible, where comprehension errors are much muddier to diagnose.
The reason nobody actually does this
Everything above is correct and almost nobody will act on it, for a boring reason: doing it properly by hand is unbearable.
To get useful data you need volume — hundreds of gap items, not thirty — and every miss tagged by type, so that “part of speech” separates from “collocation” separates from “genuinely did not know the word”. That is a spreadsheet you maintain while doing the least enjoyable exercise in TOEFL prep. Two days in, everyone stops, which is why so many people walk into this test having barely practised sixty per cent of Reading.
The useful record is not “I got 22 of 30”. It is “six of my eight misses were the wrong part of speech” — because that is a nameable, fixable habit, and “improve vocabulary” is not. Which is exactly the kind of bookkeeping worth handing to a machine.
Play Second Nature →The task as an arcade game — gaps under a clock, at volume, with every miss typed for you. Free, no card.On the numbers in this note
ETS notes that its 2026 specifications are “subject to minor revisions … until the official launch”. Not affiliated with ETS. TOEFL is a registered trademark of ETS.