TOEFL Reading 2026: A Question-Type Playbook to Hit Band 5+ on the Adaptive Section

The new TOEFL iBT condensed the Reading section, but it did not make it easier. With two adaptive passages, a hard tilt that triggers after the first set of correct answers, and roughly 35 minutes on the clock, every wrong click pulls your band down faster than it did on the old 0–30 scale. If you are sitting for a May, June, or July administration with a Band 5+ goal, you do not need more passages — you need a question-type playbook that tells you exactly what to do when each prompt appears.

This guide breaks the new Reading section into the question types you will actually face, the trap each one builds, and the move that gets you out of it on time. Read it the way you would read a chess opening repertoire: study the patterns, then drill them until your hand recognizes the position before your brain finishes reading the stem. If you are still building the broader test-taking foundation, our 8-week TOEFL iBT study plan pairs well with this playbook.

What changed, and why your old strategy will fail you

The 2026 Reading section is shorter (two passages instead of three), adaptive (your second passage is harder if you do well on the first), and graded on a 1–6 band rather than the 0–30 raw scale. That has three practical consequences nobody talks about enough.

First, every question carries more weight. On the old test, you could miss four or five questions in a section and still land in the 26–28 zone. On the new band scale, the gap between Band 4 and Band 5 is often three or four questions across the whole section. That is the difference between a “competitive” and a “qualifying” score at most graduate programs.

Second, pacing pressure is brutal in the back half. Because the second passage is harder, the questions that decide your band are also the ones with the densest vocabulary, the longest inference chains, and the trickiest “negative factual” stems. If you spend 19 minutes on Passage 1, you walk into the band-deciding passage with 16 minutes and already-tired eyes.

Third, the trap design is sharper. The new question writers know test-takers have access to ChatGPT-style summaries during prep. So the wrong answers increasingly look like “smart-sounding” summaries that are technically off by one verb or one quantifier. If you read for the gist instead of the exact claim, you will pick the elegant wrong answer every time.

The fix is a question-type playbook. Once you can name the question type in two seconds, you can execute the right move on autopilot and save your thinking budget for the genuinely hard inferences.

The seven question types you will actually face

The new Reading section recycles seven question archetypes, sometimes labeled differently across forms. Memorize the stems and the moves. Each move below is calibrated to the new ~35-minute, two-passage format.

1. Factual Information

Stem clue: “According to the passage…”, “The author states that…”, “Which of the following is true about X?”

The trap: an answer choice that is true in the real world but is not stated in the paragraph the question targets.

The move: locate the keyword from the stem in the passage, then read the sentence before and the sentence after. Your answer must paraphrase one of those three sentences. If a choice introduces a new entity, a comparison, or a cause that is not in those three sentences, eliminate it — even if you “remember” reading it elsewhere.

2. Negative Factual (“EXCEPT” or “NOT mentioned”)

Stem clue: “All of the following are mentioned EXCEPT…”, “Which is NOT discussed…”

The trap: three of the four answers feel obviously supported, and you pick the one you can’t immediately confirm — but it is actually mentioned in a clause you skimmed.

The move: in the answer column, write a tiny check next to each option you can verify in the passage. Do not try to “find the wrong one.” Find the three right ones. The unchecked one is your answer. This flips the cognitive load from search-for-absence (hard) to search-for-presence (easy) and is the single most reliable upgrade for Band 4 readers.

3. Vocabulary in Context

Stem clue: “The word ‘X’ in paragraph Y is closest in meaning to…”

The trap: the dictionary-correct synonym that does not match how the word is used in this sentence.

The move: cover the answer choices, reread the sentence with the target word, and substitute your own word. Then uncover and pick the choice closest to your substitute. This stops you from being seduced by the choice that “sounds academic” but means something different here.

4. Inference

Stem clue: “It can be inferred from paragraph X that…”, “The passage implies that…”

The trap: an answer that is a reasonable real-world conclusion but requires evidence the passage does not provide.

The move: an inference on TOEFL is one logical step away from a stated sentence — never two. Find the parent sentence in the passage, then check that the answer follows by adding only one obvious link (e.g., a definition, a contrast already set up, or a category the passage already named). If you need a second jump, it’s wrong.

5. Rhetorical Purpose

Stem clue: “Why does the author mention X?”, “The author discusses Y in order to…”

The trap: an answer that correctly describes what X is, instead of explaining why the author included X.

The move: read the sentence before X. The author’s purpose is almost always to support, illustrate, contrast with, or qualify the claim in that prior sentence. Your answer should connect X to the claim it serves — not redefine X.

6. Sentence Simplification

Stem clue: “Which of the sentences below best expresses the essential information in the highlighted sentence?”

The trap: an option that keeps the sentence’s “feel” but drops a logical connector — usually the one carrying causality or contrast.

The move: identify the connectors in the original sentence (because, although, however, despite, in order to). Your answer must preserve every connector’s logical role, even if the wording changes. Eliminate any choice that turns “because A, B” into “A and B” — that is the most common trap pattern.

7. Prose Summary

Stem clue: “Complete the summary by selecting the THREE answer choices that express the most important ideas in the passage.”

The trap: choices that are factually accurate but cover minor examples instead of the passage’s main throughlines.

The move: before looking at the choices, write one sentence per body paragraph that captures its argumentative role (claim, evidence, counter, conclusion). Your three answers should match three of those role-sentences. Anything tied to a single example, a date, or a quote is a distractor — even if it’s true.

A pacing model that survives the adaptive jump

The number-one reason Band 4 readers stall at Band 4 is that they pace Passage 1 like the only passage. They read carefully, they double-check, they finish in 21 minutes. Then Passage 2 hits with denser vocabulary and a tighter argumentative structure, and they have 14 minutes to do the questions that actually decide their band.

Use this pacing model instead, calibrated for the 2026 two-passage format:

• Passage 1: 16 minutes total. Three minutes on the passage itself, twelve minutes on the questions, one minute buffer.
• Passage 2: 18 minutes total. Four minutes on the passage, thirteen minutes on the questions, one minute buffer.
• Built-in slack: 1 minute. Hold this for the prose summary at the end of Passage 2 — it always takes longer than you think.

The key shift: spend more time on Passage 2’s text. Because vocabulary and structure are harder, an extra minute of reading the passage saves three minutes of bouncing back to find evidence during the questions. Top scorers consistently spend more reading time on the harder passage, not less.

Drill plan: four weeks to a band jump

If you have a test in roughly four weeks, this is the cycle that produces band movement.

Week 1 — Type recognition. Do one passage per day, untimed. After every question, label the question type from the seven above before checking the answer. Goal: 95% labeling accuracy. You cannot apply the move if you cannot name the question.

Week 2 — Move execution. Same one passage per day, but now timed at 16 minutes. After each question, write a one-line note on whether you executed the prescribed move. Track which moves you skipped — those are your real weak points, not the question types.

Week 3 — Adaptive simulation. Two passages back-to-back, 35 minutes total, using the pacing model above. Use a hard second passage from the official 2026 practice sets. After each session, calculate your error rate by question type. Drill the bottom two types in isolation the next morning.

Week 4 — Test-day rehearsal. Three full reading sections under exact test conditions, spaced 48 hours apart. Use the same desk, same chair, same morning routine you will use on test day. The goal is not to learn anything new — it is to make the moves automatic when nerves shrink your working memory.

The single biggest mistake high-Band 4 readers make

They reread. When they hit a question they can’t crack in 30 seconds, they go back to the passage and read whole paragraphs again. That is a Band 4 ceiling behavior.

Band 5+ readers do something different: they reread one specific clause, identified by a specific word from the question stem, and they do it within five seconds of getting stuck. If that clause does not crack the question, they make their best evidence-based guess, flag the question if the format allows, and move on. They never burn 90 seconds on a single question. The math is simple — one 90-second question costs you the time you needed for two questions on Passage 2, where the band is decided.

If you internalize one habit from this guide, internalize that one. Pick a clause, not a paragraph.

Test-day execution checklist

• Open Passage 2 with at least 18 minutes on the clock. If you don’t, your pacing failed and you need to make decisions, not careful inferences, on the remaining questions.
• Label every question’s type before reading the answer choices.
• For Negative Factual, check the three right answers, not the one wrong one.
• For Vocabulary, substitute your own word before looking at the choices.
• For Inference, accept only one logical step from the parent sentence.
• For Prose Summary, build your three role-sentences from your own notes, then match.
• Hold one minute for the final review. Do not start a fresh question with under 60 seconds left — re-examine your flagged ones instead.

The TOEFL Reading section in 2026 is not a test of how much you know. It is a test of how cleanly you can execute a small set of moves under time pressure. Learn the seven types, drill the seven moves, and respect the pacing model. That is the path from Band 4 plateau to Band 5+ — and from a “we’ll consider” to a “you’re admitted” in your fall application file.