Listening6 parts
Canadian English across accents, registers and speeds. Audio plays once.
Full-length mock exams under real test timing, all 20 scored parts, and AI feedback on Writing and Speaking scored against the four published performance standards.
Listening and Reading: 2 sets per part with no account, 5 when signed in. Writing and Speaking: the first task of every part (W1, W2, S1–S8) with a free account. Pro from CAD $9 per week.



Reviewed by the celpiptest team•Updated
The four sections always run in the same order on test day. Our mock exams mirror these durations exactly, so pacing you practise is pacing you can trust.
CELPIP-General is a computer-delivered English proficiency test accepted by IRCC for Canadian permanent residence and citizenship. It has four sections (Listening, Reading, Writing and Speaking), takes about three hours in a single sitting, and reports each section on the Canadian Language Benchmarks (CLB) scale from 1 to 12.
| Section | Scored parts | Time | Question types | Scoring |
|---|---|---|---|---|
| Listening | 6 (L1–L6) | 47–55 min | Conversations, news, discussions, viewpointsPart 5 on video | Answer key |
| Reading | 4 (R1–R4) | 55–60 min | Correspondence, diagrams, information, viewpoints | Answer key |
| Writing | 2 (W1–W2) | 53–60 min | Email, survey response | Rubric (4 criteria) |
| Speaking | 8 (S1–S8) | 15–20 min | Advice, personal experience, images, opinions | Rubric (4 criteria) |
Finish a Writing or Speaking task. You will see which of the four criteria were weak, and what to change.
Each part below has its own practice sets, vocabulary bundle and worked strategy lesson. Start with whichever one costs you the most points.
Canadian English across accents, registers and speeds. Audio plays once.
Dense texts on a hard clock. Pacing decides the band.
Both tasks AI-scored on Content, Vocabulary, Readability and Task Fulfillment.
Recorded answers against the real countdown, with rubric feedback per task.
In the real test you watch three people discuss something, and their glances, pauses and body language carry meaning the transcript does not. Most practice platforms hand you an audio file for this part, so the first video you ever see is the graded one.
Ours plays as video, once, with no replay, same as the real thing. The rest of the section stays audio, exactly as the format prescribes.

Elara
Ines
Raul
Most candidates plateau because they drill randomly. Each stage below has an exit condition, you move on when the numbers say so, not when you get bored.
Sit a single timed section under real conditions. You get a starting CLB estimate and a map of exactly which parts leak points.
Exit conditionOne full section submitted and scored.
Work the weak parts specifically (L1–L6, R1–R4, W1–W2, S1–S8), each with its own vocabulary bundle and question bank.
Exit conditionEvery weak part above your target band twice.
Full three-hour mocks, taken across sittings if you like. Writing and Speaking come back AI-scored against the four published criteria. The band is an estimate, not an official CELPIP score.
Exit conditionTwo consecutive mocks at or above target.
Short maintenance sets in the fortnight before test day, tuned to keep pacing and stamina where you built them.
Exit conditionTest day. Nothing about it surprises you.

Every section is reported as a CLB level between 1 and 12, and immigration programs usually quote a minimum per section rather than an average. That single fact should shape your whole study plan: a candidate sitting at CLB 10, 10, 10 and 7 is scored as a CLB 7 candidate by IRCC, so the weakest section is the only one worth most of your attention.
These two sections are objective. Your raw correct answers are converted to a CLB level, which means the fastest gains come from removing avoidable losses, misread question stems, running out of time in Reading Part 4, or losing the thread of a multi-speaker discussion.
On test day, certified raters score these on four criteria, and our AI returns feedback on the same four so the language you get used to is the language of the scoresheet:
Task Fulfillment is where most candidates quietly lose a band. A beautifully written email that answers two of three bullet points is capped, no matter how strong the grammar. Our Writing course drills the thirty-second outline habit that prevents exactly this.

Official results arrive within four to five business days. In our mock exams the estimate appears as soon as you submit a section, and the composite band unlocks once all four sections are in, you can spread them across several sittings, since progress autosaves.
The questions test takers ask us most often, answered in full rather than in one line.
Two prices, no tiers to compare and nothing hidden behind a sales page. Weekly billing exists because most people prepare for a fixed number of weeks.
Try the platform before you commit: no payment required.
No card. Guests get Listening/Reading previews plus 1 Writing and 1 Speaking task.
Prices in CAD · cancel anytime · the free tier stays free
One timed section is enough to see where you stand. From there the platform tells you which of the 20 parts to work on, in what order, and when you are ready for a full mock.