MET Speaking: How Much You Say Is Literally Scored

Most speaking tests judge how well you talk. Michigan's MET publishes a scale that also counts how much.

Under Task Completion, the listed bullets are relevance of response to task, quantity of language produced, and the ability to elaborate and provide relevant supporting detail.

Put that next to the other published rule, that nothing said after the timer expires is recorded, and the strategy becomes unusually concrete: fill every second of every window, and stop the instant it ends.

SpeakShark is free for exactly that kind of timed solo practice, three AI conversation sessions a day with no card, each capped at five minutes and four turns.

In this guide: no examiner in the room · the five tasks · quantity is scored · the topic thread · the persuasion task · the three categories · the visa trap · what we could not verify · how we researched this · FAQ

Key takeaways

• No live examiner. You record into your computer microphone.
• Five tasks: 60, 60, 60, 90, 90 seconds. Five minutes of speech.
• Quantity of language produced is a published scoring bullet.
• Anything after the timer is not recorded, and a countdown runs while you speak.
• The two skill MET is not accepted for Australian visas or US healthcare occupational visas.

There is no examiner

Michigan states the delivery plainly:

There is no live speaking examiner for MET, you will record your responses with your computer microphone.

The rest of the mechanics follow from that. Each task appears on the screen and is read aloud. You hear "Begin speaking now" when it is time to start. The recording begins automatically and continues for the full time allowed. When the time is up it stops and you move to the next task. A timer counts down the remaining time on screen while you talk.

Two things about that are worth absorbing before test day.

Silence is recorded as silence. In a face to face test an examiner can nudge, rephrase, or move on. Here the microphone simply runs. Nobody rescues a pause, and the pause is inside your answer.

The countdown is visible. That helps if you have practised against a clock and hurts if you have not, because a visible timer is one more thing pulling at your attention while you are trying to produce language.

This puts MET in the same delivery family as Linguaskill Speaking, which is also recorded and examiner-free, and a long way from the C1 Advanced format or our Cambridge B2 First speaking overview, where you sit with another candidate and two examiners.

The five tasks and their clocks

Between the two parts, Michigan says you are shown an introductory video to Part Two.

Five minutes of actual speech. For scale, Michigan gives the whole MET as 155 minutes, with Writing at 45, Listening at 35, Reading at 65 and Speaking at 10. So the speaking section is a small slice of the day, and the part of it where you are actually talking is smaller still. Every one of those five minutes carries weight.

Quantity of language produced

Michigan publishes the speaking scale as three categories with their sub-criteria listed. Under Task Completion:

• Relevance of response to task
• Quantity of language produced
• Ability to elaborate and provide relevant supporting detail

That middle bullet is unusual enough to be worth stating twice. On this test, the amount you produce is named as an assessed feature, not merely a side effect of being fluent.

Combine it with the recording rule Michigan also publishes:

Pay attention to the time; anything that you say after the time is up will not be recorded and will not be counted towards your score.

The two together define the tactic. A candidate who gives a tidy, complete forty second answer to a sixty second task has thrown away a third of their opportunity in the category listed first. A candidate still mid-sentence at ninety seconds loses the tail of an idea but keeps everything before it.

Practically:

1. Plan to reach the end of the window, not to finish your point.
2. Have a fallback direction for when your first idea runs out: a reason, an example, a contrast, what you would do differently.
3. Practise against a timer at both lengths, 60 and 90, until you know what they feel like.
4. Do not save the best for last. If time cuts you off, whatever you said first is what got recorded.

Twelve daily speaking drills from B1 to C1 and a 30 day plan to improve speaking both build the habit of talking to a clock rather than to a natural stopping point.

Tasks 1 to 3 are one thread

This is the structural detail that shapes Part One: all of the tasks in Part One are related to the same general topic.

Michigan's own published example makes it concrete. Task 1 says "Describe the family meal." Task 2 says "Tell me about a family meal that you really enjoyed." Task 3 says "Some people like to eat at home. Other people like to go to restaurants. What do you prefer to do? Give your opinion and reasons to support it."

One picture, then your life, then your view. The topic does not reset.

That has a useful consequence and a risk. The useful part is that vocabulary you reach for in Task 1 stays live for the next two minutes, so the words you find describing the scene are the words you will want again. The risk is that if you burn every idea you have in Task 1, Tasks 2 and 3 arrive with the well already dry.

The move is to describe the scene concretely in Task 1 and hold your opinions back, because Task 3 is where opinions are actually asked for. The B1 Preliminary extended turn is a good place to practise pure description without editorialising, since describing one photograph is the whole task there.

Task 5 asks you to persuade a person

Task 4 and Task 5 both run 90 seconds, and Task 5 is the one with a shape you rarely see in an English exam.

Michigan's published example: "A local school is considering requiring all students to attend school on Saturdays. Many people think this is unfair to the students. I am the school principal. Tell me what you think about this and try to convince me to agree with you."

You are given a role to address and asked to change that person's mind. That is closer to a workplace situation than to an exam answer, and it rewards different language: acknowledging the other position, conceding a point, then pressing an argument.

Task 4, by contrast, asks for both sides without asking you to pick a winner: given a situation, "What are the advantages and disadvantages of taking this job?" So Part Two runs balance first, then persuasion.

Role play scenarios for speaking practice is the closest drill for the persuasion task, and C1 Advanced Part 4 covers justifying an opinion at length, which is the underlying skill both tasks share.

The three scoring categories

Michigan names them and lists what sits under each:

Michigan states that MET Speaking is scored by trained raters in these three categories.

Two observations. Pronunciation is one of four items inside one of three categories, so it carries a fraction of a third of the judgement. That matches what Cambridge's published examiner data shows for its own exams, where pronunciation is the criterion candidates score highest on rather than lowest. If you are spending most of your preparation on accent, the published scales at two different exam boards both suggest that time is going to the wrong place.

Hesitation is named separately from fluency. Those are related but not identical, and hesitation is the one that responds fastest to simply speaking more often. How to improve English pronunciation and six pronunciation fixes in fourteen days cover clarity, and why your English speaking is not improving covers why silent study leaves hesitation untouched.

The two skill version does not work for visas

MET comes in a four skill and a two skill version, and the difference is consequential.

Michigan states that the two skill MET "can only be taken in an Authorized Test Center" and is "used by some institutions to meet internal or local requirements". Then two specific warnings:

If you are a healthcare professional applying for an occupational visa to work in the U.S. you will need a 4-skill exam taken in an approved test center.

If you plan to apply for an Australian visa for any purpose, you will also need a 4-skill MET taken in an approved test center.

Both conditions have two halves: four skills and in an approved test center. Sitting the right version at home would fail the second half.

Michigan places MET at CEFR A2 to C1, digital, on demand, available at test centers and at home in some countries, and states that more than 2,000 organisations recognise it for immigration, work and study, including recognition by the Australian Department of Home Affairs.

That A2 to C1 span is wider than most single exams attempt. The Council of Europe's global scale puts A2 at communicating "in simple and routine tasks requiring a simple and direct exchange of information", and C1 at expressing yourself "fluently and spontaneously without much obvious searching for expressions". A five task section reporting across that whole distance is why the tasks climb from describing a picture to persuading a named person. A free daily session is a reasonable way to keep moving along that span between now and test day. If Australia is your destination, the English test for an Australian visa covers what the department accepts more broadly, and which English test for a UK visa covers the British route, which works differently again.

Always confirm the current requirement with the body that will read your result. Recognition changes, and a blog is not the authority on your application.

What we could not verify

We have not reproduced the MET Speaking rating scale band descriptors. Michigan publishes the numbered levels in its resource pack and those are the authority. We have named the three categories and the sub-criteria Michigan lists under each, and stopped there.

We could not confirm the timing of the Part Two introductory video, only that Michigan says one is shown between the parts.

We did not verify the two skill version's task list. Everything above describes the speaking section, which by definition belongs to the four skill exam.

We are not stating fees, score requirements or pass marks. Requirements belong to the receiving organisation.

We did not sit the test. All of this comes from Michigan's published material.

How we researched this guide

The delivery description, the five tasks, the 60 and 90 second timings, the example prompts, the rule about speech after the timer, the three scoring categories and their sub-criteria all come from the 2026 MET Speaking Resource Pack, a free PDF published by Michigan Language Assessment, downloaded and read in full.

The 155 minute total, the per-section timings, the CEFR A2 to C1 range, the digital on demand format, the 2,000 organisation figure and the Australian Department of Home Affairs recognition come from Michigan's MET page. The four skill versus two skill distinction and both visa conditions come from Michigan's MET preparation page.

Two process notes worth recording. Our first attempt at these sources guessed the URLs and every one returned 403 or 404. The real paths sit under a differently named section of the site, and they were found by navigating rather than by guessing. We now treat any exam source URL we have not clicked as unverified.

Second, the two Michigan HTML pages linked below return 403 to an automated request while serving normally in a browser, which is ordinary bot filtering rather than a broken link. Both were read on screen. The resource pack PDF downloads without any such block.

Practise speaking, from SpeakShark

SpeakShark is an AI English speaking practice app, and it is our pick for MET because the test is solo, recorded and timed, and so is the practice. There is no partner to arrange. You speak into a microphone, something answers, and you keep going until the window closes. You talk, the AI answers what you actually said, and you get speaking feedback while the conversation is still going. The free tier gives basic feedback; the detailed pronunciation and grammar breakdown is on Premium.

Being straight about the limits: a free session runs five minutes with four turns, which suits daily practice of the individual tasks rather than simulating the full section.

Start free with three sessions a day and no card. Paid sessions run ten minutes with unlimited turns. Limits are on the pricing page, and how it works walks through a full session.

We are a speaking improvement tool. We are not an exam preparation provider, and we are not affiliated with Cambridge English, the IELTS partners, Pearson or any exam board. For test format, booking and official practice material, go to the exam body directly. Use SpeakShark to make your spoken English stronger, and use official material to learn the test.

• Site: speakshark.com
• Support: speaksharksupport@gmail.com
• Free tier: 3 AI conversation sessions per day, no card

Sources

• Michigan Language Assessment, MET Speaking Resource Pack 2026: the no-examiner delivery, all five tasks with their timings, the published example prompts, the rule that speech after the timer is not recorded, and the three scoring categories with their sub-criteria
• Michigan Language Assessment, MET: the CEFR A2 to C1 range, the 155 minute duration, the per-section timings, the digital on demand format and the recognition claims
• Michigan Language Assessment, MET preparation: the four skill versus two skill distinction and both visa conditions
• Council of Europe, CEFR Table 1, common reference levels global scale: the A2 and C1 descriptors that bracket the range MET reports against

FAQ