How long is the TOEFL Speaking section in 2026?

The specification lists an estimated time of about 8 minutes for the Speaking section, within a total listed test time of 1 hour 23 minutes to 1 hour 29 minutes. Listed values are subject to ETS’s stated pre-launch revisions.

Is the Speaking section adaptive?

No. The 2026 Reading and Listening sections use multistage adaptive designs; Speaking and Writing are linear.

Are the old independent and integrated speaking tasks still on the test?

The current specification lists two Speaking task types — Listen and Repeat (7 items) and Take an Interview (4 items). A guide built around long prepared opinion speeches with reading and listening sources describes an earlier format.

Who scores my speaking responses?

Per the specification’s item-type notes, Speaking tasks are AI-scored constructed responses. Practice scores from any provider — including Reach120 — are estimates, never official results; only ETS issues official scores, on the 1–6 half-band scale.

Current format, sourced

The 2026 TOEFL^® Speaking section: two task types, eleven short items

The current TOEFL^® test Speaking section lists 11 short items in two task types — Listen and Repeat (7 items) and Take an Interview (4 items) — in a linear design, at roughly eight minutes listed time. There are no long prepared speeches; each response is short, spoken immediately, and scored from your recorded speech.

What the specification lists

ETS states — from the blueprint table of the 2026 test specifications, accessed 17 July 2026.

Listed field	Listed value
Task types	Listen and Repeat (7 items); Take an Interview (4 items)
Items	11
Maximum raw points	55
Estimated time	~8 min
Design	Linear — Speaking and Writing are not adaptive

These are the published blueprint’s listed values, which ETS marks as subject to minor revision until the official launch.
The listed total of 120 items across the whole test includes non-scored Reading and Listening items. Speaking has none of its own.
All Speaking tasks are AI-scored constructed responses, per the specification’s item-type notes. That is the opposite end of the scale from the selected responses in Reading and Listening — and machine-scored is not exclusive to those two sections: the Writing section’s ten Build a Sentence items are machine-scored too.

The two task types, and how to practice each

Task names and item counts are the specification’s. The description of the skill measured and the practice pattern are Reach120’s reading, not ETS’s preparation advice.

Listen and Repeat

7 listed items

You hear a sentence and repeat it aloud, reproducing what was said. The measured skill is real-time phonological control: holding a full sentence in working memory, then producing it with intelligible pronunciation, natural stress, and intact grammar — under the clock, with no text to read from.

Practice pattern: Work in short daily reps, not long sessions. Shadow sentences slightly above your comfort length; when a repetition collapses it usually collapses at a specific point — an unstressed function word, a consonant cluster, a clause boundary. Drill that point, not the whole sentence. Record yourself and compare rhythm, not just words.

Take an Interview

4 listed items

You answer spoken interview-style questions with short, direct spoken responses. The measured skill is on-demand coherent speech: answering the question actually asked, supporting it with something concrete, and finishing inside the time — not delivering a memorized template.

Practice pattern: Train a reusable answer shape — direct answer, then one concrete example, then one sentence of outcome or reason — until it is automatic, then vary topics daily. Timed reps matter more than perfect sentences: an unfinished elaborate answer usually communicates less than a completed simple one.

One prompt, one outline

Reach120-original worked example. This interview prompt was written by Reach120 to illustrate the task pattern. It is not an ETS item and reproduces no official material.

Many people prefer studying with background noise, like a café. Others need silence. Which works better for you, and why?

1
Direct answer first — one sentence
“For me, a quiet room works much better.”
2
One concrete reason or example — two to three sentences
“When I studied for my biology final in the library café, I reread the same page three times without remembering it. In a silent room I can finish the same reading in half the time.”
3
Close with the consequence — one sentence
“So now I only review in quiet places and save the café for breaks.”

What a rater listens for

Reach120 interprets — no ledger row covers a published Speaking rubric, so treat this as our reading of what constructed-response speech scoring generally rewards, not an ETS criteria list.

Can the response be understood without effort — intelligibility and pacing?
Does it answer the question that was asked, not a rehearsed neighbor of it?
Do the sentences connect (“when… so now…”) rather than sit as fragments?
Is it complete within the time?

A short, finished, on-topic answer beats an ambitious unfinished one on every one of those dimensions.

What a practice score on this section means

All Speaking tasks are AI-scored constructed responses in the specification, and any speaking score a practice tool produces — including Reach120’s — is a practice estimate, never an official outcome. Official score reports for tests taken on or after January 21, 2026 use a 1–6 scale in half-band increments as the primary official scale, with a comparable 0–120 estimate also shown during ETS’s transition through January 2028. No practice number here is that scale.

This page promises no live Speaking practice. Any future speaking surface covers at most the two listed task types, with coverage disclosed as partial, and no practice result combines into a readiness number for the whole test.

Common questions

How long is the TOEFL Speaking section in 2026?: The specification lists an estimated time of about 8 minutes for the Speaking section, within a total listed test time of 1 hour 23 minutes to 1 hour 29 minutes. Listed values are subject to ETS’s stated pre-launch revisions.
Is the Speaking section adaptive?: No. The 2026 Reading and Listening sections use multistage adaptive designs; Speaking and Writing are linear.
Are the old independent and integrated speaking tasks still on the test?: The current specification lists two Speaking task types — Listen and Repeat (7 items) and Take an Interview (4 items). A guide built around long prepared opinion speeches with reading and listening sources describes an earlier format.
Who scores my speaking responses?: Per the specification’s item-type notes, Speaking tasks are AI-scored constructed responses. Practice scores from any provider — including Reach120 — are estimates, never official results; only ETS issues official scores, on the 1–6 half-band scale.

What this page does not tell you

This restates the blueprint’s listed Speaking values, which ETS marks as subject to minor revision until official launch, plus one Reach120-original example. It does not state per-item timing, rubric wording, or scoring conversions ETS has not published; the rater notes above are labeled interpretation; and a practice result is not an official score.

Audit your speaking material against this section The whole-test map

Sourced from the ETS 2026 test specifications, accessed 17 July 2026. Next review 18 August 2026, or immediately on any ETS specification change.

The 2026 TOEFL® Speaking section: two task types, eleven short items