CELPIP Listening Mock Test 6

Mock Test 6 is the hard one. It carries what every listening test on this site carries, six parts and 38 scored questions in the 2026 format, plus an unscored practice task, and then it does something the others do not: it makes people say a thing and then take it back.

That is the whole design. On an easier test you catch a fact. Here you have to catch a fact being revised, and notice which version survived. This is the test you sit to find out whether you are ready, so take it after the others rather than instead of them, and read the tips below before you do.

What is on CELPIP Listening Mock Test 6?

All six clips are set in Mississauga. Each one is listed here with the thing that makes it hard, because on this test those are the same thing.

Why this one is the hard test

The blurb on this test says the speakers change their minds and the answer is rarely in one place. That is not marketing, it is the specification it was written to, and it shows up in all six parts.

In Part 1, Wanda works out that a temp is short a hundred and eighty dollars because he has two live badges and payroll only pays what it can see under one number. She tells him to go to the agency trailer. Then she asks when exactly he moved to days, hears the seventeenth of March, and takes it straight back: anything after the first of March is the site's system, not the agency's, so it is the office upstairs and the door on the left. A candidate who wrote down the trailer and stopped listening has an answer the audio has already overruled.

Part 3 does it again, and colder. The registrar tells the woman there are two roads, that the second is better, and that most people in her position never even hear about it: no classes, a portfolio and an exam, and free for a mature student. She says that is the one, obviously. Then he asks when she needs it on paper, hears the fifth of December, and scratches it, because the challenge only runs in October and April and the results do not post until mid December. It is the right road and the wrong month, and the credit she actually ends up with is the accelerated course starting Thursday.

Part 4 corrects a number inside a news report: the figure released last week put the desk at six hundred thousand a year, and the Centre has since corrected that to five hundred and fifty. Part 5 has Gus arguing hard for online check-in, hearing that nine of twenty-two patients gave a number that is not a mobile, and saying out loud that he is going to walk that back, partly. And Part 6 opens with a land economist telling a planning committee that his own March report was wrong on a material point and he would rather say so there than have it said for him at the Tribunal.

Part 2 is the exception that proves it. Nobody reverses anything, because nobody commits to anything: Trevor will not say he is running for the condo board and Doreen will not stop asking. The last thing he says is that it is not a no. The questions ask you where he landed anyway, and the evidence is entirely in how he says it, which is what an inference question is for.

CELPIP listening tips and tricks for the hard test

General listening advice is written for clips where the answer sits still. These are the habits that pay when it does not, and each one is aimed at something that actually happens on Test 6.

• Leave room for the second answer. This is the tip for this test. When a speaker gives you a fix, do not close the line: leave a gap under it. Wanda's trailer becomes the office upstairs. The registrar's prior learning assessment becomes the accelerated course. If your note has one answer per question, a reversal has nowhere to go and you will keep the version the audio threw away.
• Mine the introductory statement. You are told before Part 6 that a land economist is addressing a planning committee about a housing application. That is the difference between hearing "residual land value" cold and hearing it inside a frame you already built. It costs nothing and it is handed to you every time.
• Set the notes up before the audio, not during it. Once a clip starts you cannot pause it, rewind it or hear it twice, so anything you can write in advance is free. Three columns for Part 5. A line for the news item's name. Space for a number to be corrected.
• Label Part 5's speakers by position. Estelle on the left, Emeka in the centre, Gus on the right, labelled as they are introduced. Three voices arguing is where notes stop being optional, and the one who moves is Gus.
• Take the last name seriously in Part 4. She is Miriam Vachon once and Vachon thereafter, which is how news copy works everywhere. If your ear is waiting for the full name again you will miss what she said about spending the summer writing letters.
• Answer the confident ones first, but only in Parts 4 to 6. That is where every question sits on one screen with a single budget and no fixed order, so sweeping the certain ones first and leaving the rest the remainder actually works. In Parts 1 to 3 there is nothing to reorder: the question is spoken to you, one at a time, thirty seconds each, and the test moves on at zero whether or not you did.
• Listen to what you understood, not to what you missed. Part 6 will hand you a pro forma, a residual land value and a ninety-nine year encumbrance. You are not scored on that vocabulary. You are scored on whether you caught that he changed his mind about the land cost and why. A word you lost is one word. A minute spent grieving it is the rest of the clip.
• Never leave a blank. One point per correct answer, nothing deducted for a wrong one, so a blank and a wrong answer are worth exactly the same and a guess is free. There may also be one unscored part you are never told the identity of, which is a further argument against writing off a part that is going badly.

How the section runs, and what to expect from a hard set

Listening is the first section of CELPIP, so this is the one you meet before your ear has warmed up on anything. The six parts carry 41 minutes of clock and Paragon schedules roughly 47 to 55 minutes for the section including its instruction screens. Difficulty is designed to climb from Part 1 to Part 6, so the three minute talk at the end is meant to be the hardest thing you hear, and on this test it is.

The rules that bite hardest here are the ones about the audio. Every clip starts on its own and plays a single time. There is no pause, no rewind, and no progress bar creeping along to tell you how much is left to come. On a set where a speaker withdraws an answer ninety seconds in, that is the entire difficulty: you cannot go back for the sentence you only half caught, so the note you made the first time is the only copy of it you will ever have.

The question formats then split the section in two, and on a hard set that split is worth real marks. Parts 1 to 3 speak each question to you, one at a time, thirty seconds apiece, in an order you are not allowed to leave. Parts 4 to 6 print them instead: sentence completions with a gap to fill, every question in the part sitting on one screen, a single budget for the whole screen, and any order you like. That is why the tip above about taking the confident ones first pays in the second half of the section and means nothing in the first.

The result is marked, not judged. Every question has one correct answer, it is checked against the key on your own machine the second you submit, and there is no evaluation and nothing to wait for. Each item then carries a sentence naming what in the audio makes the key correct, and review mode replays the clips you only heard once with the transcript beside them. On a test built around reversals that matters more than usual: it lets you go back and find the sentence where the answer changed, which is a different lesson from being told you got it wrong. Your raw score out of 38 converts on the CELPIP listening score chart.

A last word on the score itself. This test is harder than the exam needs to be at most band targets, so a lower number here is not the number you would get on test day. What it is good for is finding the specific habit that costs you marks when the audio stops cooperating, which is exactly the habit a comfortable test will never expose.

The two things we do not pretend about

Nobody in these clips has a Canadian accent. Paragon's own audio does, and what you will hear here is North American instead, because there is no Canadian synthetic voice on any engine we can reach: the one that reads these clips has 102 voices and not a single Canadian among them. So we say North American and mean it. The Canadian content is not a claim, it is in the scripts: a lead hand and a temp on a Mississauga distribution floor, a hydro reconciliation, a PR card, a health card, condo board nominations closing on the fourteenth, Peel and Streetsville and the Dundas Street corridor, Canadian spelling throughout. In a synthesized voice the distance between Standard Canadian English and General American is small, and we would rather tell you than let you find out.

The real test plays a video in Part 5. It is the only one in the section, and here it is audio with the three speakers shown and labelled by position. Paragon's own strategy notes say that no question tests a visual detail, which is what makes the swap fair rather than convenient. The thing the video genuinely gives you is the ability to label three speakers by where they are and keep them straight, and that is preserved exactly.