Standardized English testing just changed — again.
The Enhanced TOEFL iBT, rolling into full use by 2026, is unlike any previous version. ETS calls it a transformation. I call it something else entirely: the TOEFL IRT — the In Real Time TOEFL.

Because that’s what this new format really measures: not memory, not templates — but real-time speaking performance.

From iBT to IRT: The Shift That Redefines TOEFL Speaking

For nearly two decades, the TOEFL iBT (Internet-Based Test) leaned on integration: read, listen, prepare, then speak. In the Enhanced TOEFL 2026, that model is gone. There’s no reading passage, no listening clip, no note-taking, and no prep time. You listen, think, and respond — live.

That’s not “internet-based.” That’s real-time English.

How TOEFL iBT (Legacy) and Enhanced TOEFL (2026) Differ

Feature	Legacy TOEFL iBT	Enhanced TOEFL 2026 (TOEFL IRT)
Task Format	4 tasks: 1 Independent + 3 Integrated	2 tasks: Listen & Repeat + Interview
Preparation Time	15–30 seconds per task	None — responses begin immediately
Scoring System	Human + AI (hybrid)	AI-only — full automation
Assessment Focus	Integration of reading, listening, and speaking	Real-time fluency, pronunciation, and spontaneity
Timing	~17 minutes	~8 minutes
Key Skills Measured	Comprehension, summarization, template structure	Delivery, Intelligibility, Language Use, Topic Development

The Real Transformation: Measuring Performance, Not Recall

The old Speaking section measured how well you could assemble information.
The new test measures how well you can perform in English.

When you speak on the new TOEFL, nothing you say is graded by someone’s opinion. It’s graded by AI. Every pause, vowel, and word boundary is measured. The system isn’t listening for ideas — it’s listening for performance: acoustic, linguistic, and rhythmic.

What TOEFL Speaking AI Tracks

Metric	Description	Impact on Score
Speaking Rate	Average words per second across the response	Too slow or too fast lowers Delivery; stability wins
Pause Frequency	Number of silences per 10 seconds	More pauses = lower Delivery; controlled micro-pauses are fine
Pronunciation Accuracy	Percentage of words recognized cleanly by the model	Primary driver of Intelligibility
Pitch Range / Stress	Variation in intonation and stress-timing	Natural prosody boosts Delivery and Intelligibility
Response Length	How fully you use the time window	Completeness and control (e.g., ~45s per Interview item)

These feed four constructs: Delivery, Intelligibility, Language Use, Topic Development — the basis of your TOEFL Speaking score out of 30.

Why “In Real Time” Is a Better Name

“iBT” (internet-based) made sense in 2005. In 2026, internet delivery is assumed. What’s new is the measurement philosophy:

No time to rehearse
No templates to lean on
No human rater to interpret your ideas

Everything is captured and scored in real time by AI. That’s why TOEFL IRT describes it better. It’s not testing your knowledge; it’s testing your control.

The End of “Perfect Accent” Anxiety

Accent doesn’t lower your TOEFL Speaking score. Poor intelligibility does. The system measures clarity and rhythm, not where you’re from. You can keep your accent and still earn a perfect 30 — if your pronunciation is consistent and your rhythm is stable.

Preparing for the TOEFL IRT Era

If the old TOEFL rewarded accuracy under prep time, the new one rewards stability under pressure. Train the way you’ll be scored: with timing, rhythm, and repeatable structure.

What Your TOEFL Speaking Practice Should Include

Practice Element	Goal	Method
Pronunciation Drills	Improve Delivery & Intelligibility	Daily 10-minute sound–stress–sentence routine
Timed SPER Responses	Build structure under pressure	45-second bursts using random prompts (Start–Point–Example–Restate)
Fluency Tracking	Quantify progress	Review Speaking Rate, Pause Frequency, and Pronunciation Accuracy
Mock Interviews	Simulate test rhythm	4 questions back-to-back, no pauses or notes
Playback Analysis	Spot hesitation points	Listen for rhythm breaks, not just mistakes

Real preparation now means data, discipline, and control — not memorization.

Final Thought

ETS doesn’t call it the TOEFL IRT, but they might as well. The Enhanced TOEFL iBT is a new category: it measures real-time English performance. Templates are useless here. Rhythm, clarity, and composure win.

Welcome to the era of In Real Time English.

Frequently Asked Questions

Question	Answer
What’s different about the new TOEFL Speaking section?	Two tasks: Listen & Repeat (7 items) and Interview (4 items). No prep time, no reading, no notes — you respond instantly, simulating spontaneous conversation.
How are TOEFL Speaking scores calculated?	Fully by AI. The engine measures Delivery, Intelligibility, Language Use, and Topic Development using acoustic and linguistic features, then scales to the 0–30 TOEFL Speaking score.
Does accent affect my score?	No. Accent doesn’t lower your score; poor intelligibility does. Stable pronunciation and predictable rhythm are what matter.
How should I prepare for the Enhanced TOEFL 2026 Speaking section?	Train performance: timed 45-second answers, daily pronunciation control, SPER structure, and post-practice data review. Use a TOEFL Speaking practice test workflow and analyze results after every session.
What counts as a strong TOEFL Speaking score now?	25+ typically reflects solid fluency and clarity. Consistency across all four constructs — not perfection — drives high totals.

Why the New TOEFL iBT Should Be Called TOEFL IRT: Speaking in Real Time