TOEFL iBT (2026 Format): Complete Study Guide

The TOEFL iBT format that launched on 21 January 2026 is the current test. This guide covers the new section order, task types, multistage adaptive Reading and Listening, the 1–6 score scale, traps, and a study plan — without inventing score guarantees.

Product truth: Speaking and Writing are scored against a rubric. Old-format practice (pre-2026 task mix and 0–120-only framing) will not predict your new sitting. Train the tasks you will actually face.

Exam Overview (2026 format)#

Section map#

---

What Changed in 2026#

Why the change matters for measurement#

Old-format raw percentages do not map cleanly onto the new adaptive Reading/Listening stages or the new productive tasks. A sitting on Acedify that uses 2026 task names and clocks is evidence for this test. A legacy mock is evidence for a test you will not take.

---

Reading (~30 min, up to 50 items)#

Reading is multistage adaptive: performance on an early stage routes you into a later stage with a different difficulty profile. Your job is still the same — accurate reading under the clock — but early misses can change what you see next.

Task types#

Adaptive reality (without superstition)#

• You cannot “game” the router by slowing down on purpose — the clock still runs.
• Treat every item as scored evidence for this sitting.
• After practice, ask: which task family is the holdback — Complete the Words, Daily Life, or Academic Passage?

Reading tactics#

1. Read the question stem before diving deep on multi-item academic sets — know the target.
2. Complete the Words: fit grammar and meaning; a plausible word that breaks the sentence is wrong.
3. Daily Life: respect the document type (email, notice, schedule) — purpose beats trivia.
4. Academic Passage: wrong options often reuse passage words in the wrong relationship.

Complete the Words — deciding line#

Work under time: assemble candidates quickly, then verify the whole sentence reads as English — not as a bag of plausible tokens.

Read in Daily Life vs Academic Passage#

---

Listening (~29 min, up to 47 items)#

Listening is also multistage adaptive. Audio still obeys a one-pass discipline — you do not get a free rewind because you “almost had it.”

Task types#

Pacing buckets in Listening terms#

Listening tactics#

1. Preview when the interface allows — know what you must listen for.
2. Listen for function on Choose a Response — what does the situation require next?
3. On Academic Talk, track structure (problem → evidence → implication), not only keywords.
4. Reset between items — adaptive stages punish lingering on a lost item.

Task-family drill map#

Do not only re-listen to the same clip until it feels familiar. Familiarity is not transfer. Transfer is a new clip, one pass, same task family, cleaner buckets.

---

Writing (~23 min, 12 tasks)#

Writing is linear (not adaptive) and mixes short constructed tasks with longer productive tasks.

Task mix#

What “scored against a rubric” means here#

For Email and Academic Discussion, raters (human and/or AI under ETS design) look for task fulfillment, organization, language control, and appropriateness — not “sounds fancy.” Build a Sentence is closer to form-focused accuracy: the sentence must work.

Writing tactics#

1. Build a Sentence: lock grammar relations (subject–verb, modifiers) before polishing vocabulary.
2. Email: identify audience and purpose in the first seconds; cover every required content point.
3. Academic Discussion: respond to the discussion, not to an imaginary standalone prompt. Acknowledge a point, then add yours with a reason.
4. Clock share: do not burn the section on one perfect email while Build a Sentence items remain untouched.

Email checklist (Task Achievement first)#

Academic Discussion checklist#

---

Speaking (~8 min, 11 tasks)#

Speaking is short, dense, and redesigned. Product framing: Listen and Repeat (7) + Take an Interview (4 items) — with Interview items framed as five points each in Acedify’s product language for how the task is taught and reviewed.

Listen and Repeat (7)#

Tactics:

• Catch chunking and stress the first time — there is no rehearsal loop on test day.
• Do not “improve” the sentence into a paraphrase; the task is reproduction.
• If you blank, recover the meaning units you have rather than freezing for a perfect replay that will not come.

Take an Interview (4)#

In Acedify’s product framing, each Interview item is reviewed across five points (content coverage and delivery dimensions used in coaching) — use that breakdown to name gaps after a sitting, not as an official ETS scorecard.

Tactics:

• Answer the question asked — do not deliver a memorized speech for a different prompt.
• Open with a direct stance, then one or two supporting points.
• Prefer clear structure over rare vocabulary that collapses mid-sentence.

Speaking criteria mindset (productive skills)#

Interview: five-point review habit (product framing)#

After each Interview response on Acedify, review along five coaching points such as:

1. Direct answer — did you address the question in the first seconds?
2. Development — reason or example beyond a single clause?
3. Coherence — could a listener outline your path?
4. Language accuracy — did errors block meaning?
5. Delivery — pace and clarity under the time box?

That five-point habit is for naming gaps between sittings. It is not a claim about ETS’s internal scoring formula.

Listen and Repeat vs Interview — do not cross-train blindly#

Balance the week to the 7 + 4 mix you will actually sit.

---

Scoring: 1–6 vs Old 0–120#

What a useful practice range reports#

• Section-level signal on the 1–6 framing (not a fake “guaranteed ETS score”)
• Which section holds the estimate
• Pacing buckets on adaptive Reading/Listening and rubric notes on Writing/Speaking

Rough mindset (not a conversion table)#

ETS publishes official concordance / comparable scores during the transition. Do not invent your own “3.5 = 80” shortcuts from a blog. Use the score report your universities ask for, and practice the 2026 tasks that feed that report.

---

Common Traps#

---

Study Plan (4 weeks, no score guarantees)#

Aimed at the 2026 constraints. No invented “+1.0 band in two weeks” claims — ranges move when evidence accumulates under real task names and clocks.

Week 1 — Learn the new language of the exam#

• Map all task names: Reading (3 families), Listening (4), Writing (3 types / 12 tasks), Speaking (2 types / 11 tasks).
• Untimed samples of each; study the deciding line on misses.
• One timed Reading stage set + one timed Listening stage set (one-pass audio).
• One Writing mix: several Build a Sentence + one Email draft against a checklist.
• Speaking: Listen and Repeat set + one Interview item recorded.

Week 2 — Drill the holdback#

• If Complete the Words or Academic Passage holds Reading, live there between sittings.
• If Choose a Response or Academic Talk holds Listening, drill that family, then sit timed again.
• Writing: alternate Build a Sentence accuracy blocks with Email / Academic Discussion rubric reviews.
• Speaking: Interview responses marked on task fulfillment + intelligibility; separate Listen and Repeat accuracy drills.

Week 3 — Full 2026 constraints#

• Combined Reading + Listening under adaptive-style timing.
• Full Writing (~23 min) with all task types represented.
• Full Speaking (~8 min) continuous.
• After each sitting: fluent / fragile / guessed / gap. Drill only what moved (or failed to move) the estimate.

Week 4 — Narrow the range#

• Two more new-format sittings with recovery between them.
• Light error-log review — no legacy-only cramming.
• Confirm the score scale your institutions want (1–6 and/or transitional 0–120).
• Rest. A narrowing range beats one lucky legacy mock treated as destiny.

---

Know your range before test day#

Format charts help. A timed 2026-format sitting shows whether adaptive Reading, one-pass Listening, Build a Sentence, Email, Academic Discussion, Listen and Repeat, or Interview is holding the estimate.

• Timed diagnostic with section-level signal on the current format
• Speaking/Writing reviewed against a rubric (diagnostic — not an official TOEFL score)
• Pacing buckets: fluent, fragile, guessed, gap
• Practice free; timed sittings and readiness as a range on Pro

Browse the TOEFL hub →

---

Next step: Start practicing free on the TOEFL hub. Sit the 2026 task mix under the clock — then read what the range says before you book.