IELTS guides: every part of the test, explained

An average of 6.25 becomes an overall 6.5 — but 6.125 becomes 6.0. The official rounding rules, with ielts.org's own worked examples, and why one half band in one skill can decide it.

The official conversion tables that exist, the caveat attached to them, and why the same 30/40 in Reading is Band 7 on Academic but Band 6 on General Training.

Listening and Speaking are identical on both tests. But GT Reading needs more correct answers for the same band — 30/40 is a band 7 on one paper and a 6 on the other, per official tables.

ielts.org announced paper IELTS ends mid-2026 — yet official pages still sell it with 13-day results. What's actually true, dated claim by dated claim, and what still differs on computer.

You can now retake one IELTS skill instead of all four — but canada.ca says flatly that Express Entry won't take it. Read who accepts it before you pay.

Your Test Report Form carries no expiry date. The famous two years is a recommendation from the IELTS partners — and a hard requirement from IRCC. Knowing whose rule it is changes your planning.

Most IELTS fee tables online carry no date — some are years old. Every figure here was checked 24 Aug 2026, with the three booking mistakes that cost far more than the fee itself.

IRCC accepts both, so 'which is easier' is the wrong question — both test makers deny it exists. The real difference: IELTS Speaking is a human interview, CELPIP is a microphone.

Bring the wrong ID and you don't test. Keep your phone and you're disqualified. And no, you can't wear a watch — even an analogue one. The official rules, quoted.

The IELTS partners give away the actual examiner criteria PDFs and full past-paper mocks — free. Here is the complete index, what each item is genuinely best for, and the one thing none of it can do.

IELTS puts a human examiner across the table — PTE scores you with 'AI and human expertise'. Neither is easier, but one PTE version is useless for Express Entry. Which one?

ETS moved TOEFL iBT to a 1–6 scale in January 2026 — most comparison pages haven't noticed. What actually differs now: scales, delivery, results speed, and who accepts what.

Most IELTS plans open with 'week 1: grammar' before knowing which skill is losing marks. The band arithmetic says a targeted half-band beats a month of general study — measure first.

Three UKVI variants exist, and every difference the official pages actually state is about where and how you sit the test. What is verified, what is not, and the booking mistake that wastes a fee.

IRCC's own IELTS-to-CLB chart is not flat — CLB 8 wants Listening 7.5 but only 6.5 elsewhere. If you're planning around an overall band, you're reading the wrong number.

The exact criteria examiners use for Writing and Speaking are free PDFs on ielts.org. Here is where they live, how to read a band row, and the version trap.

A 7 is roughly 30 raw marks in Listening, 30 in Academic Reading, 35 in GT Reading — and in Writing and Speaking it is specific descriptor sentences. Four different walls, one number.

Yes, there is an official IELTS you can sit at home. Its own official page says it is not accepted for immigration — and most people find out after paying.

A Part 1 chat about travel arrangements pays exactly the same as a Part 4 academic lecture — one mark each, heard once. Know which marks are cheap before the audio starts.

You can hear the answer perfectly, write it down accurately — and still lose the mark. Three words where two were allowed, one wrong letter. Here is how each Listening type takes marks off you.

Seven raw marks separate Band 6 from Band 7 on the official table — and some of yours are dying to spelling, word limits and transfer slips, not to comprehension. A four-part diagnostic.

Part 1 is the easiest audio you will hear all test — and the part where correct listeners lose marks anyway, because a name heard perfectly but spelt wrongly scores zero.

Part 2 is officially a monologue — one voice, no dialogue turns to rescue you when you drift. Here's why maps punish that, with a worked example and the direction-language traps.

Part 3 rarely asks what was said — it asks who ended up believing it. Speakers agree, disagree and change their minds, and every wrong option quotes the recording accurately.

Part 4 is officially a monologue on an academic subject — one voice, played once, and worth exactly one mark per question, same as Part 1's phone call. Here is why people lose marks there anyway.

IELTS says it plainly — you lose marks for incorrect spelling. Dates, prices and spelled-out surnames are decided by transcription skill, not English level, and you can drill it in a week.

The official timing line most candidates misread: 60 minutes including transfer time. There is no extra window to copy answers — here is what that does to your pacing.

FALSE means the text contradicts the statement. NOT GIVEN means the text is silent. Every trap in this question type lives on that line — walk through four statements that prove it.

The eleven official Reading question types are not eleven flavours of the same task — each rewards a different eye movement, and practising one does not train the others.

40 questions, 60 minutes, and officially no extra transfer time — 90 seconds a question before you have read a word. Four timed failure modes lose more Reading marks than vocabulary ever does.

Y/N/NG looks like True/False/Not Given in different letters. Officially it tests something else — the writer's views, not the text's facts — and that changes how you answer.

The wrong heading usually shares words with the paragraph. The right one shares its job. A worked example with all four options torn down — the trap, the detail, the theme, the answer.

The summary never reuses the passage's words — so scanning for them finds the wrong paragraph. How to match meaning instead, and the official word-limit rule that deletes right answers.

Matching information asks where a fact lives — not what a paragraph is about. Use a matching-headings strategy here and you will hand the test exactly the wrong paragraph.

The unfinished sentence already tells you the answer's word class — a verb, a number, a noun phrase. Use the grammar first and half the passage stops being a candidate.

IELTS Reading distractors follow repeatable shapes — true-but-not-asked, half-right, passage words with flipped meaning. Learn to name each shape and elimination stops being guesswork.

GT Reading gives you notices, adverts and workplace documents — then asks for more correct answers than Academic for the same band. 30/40 is Band 7 on Academic and Band 6 on GT. Here's why.

A flow-chart's boxes follow the process's order — not the paragraph's. Here's the mismatch that costs marks, one worked example torn down box by box, and the word-limit rule quoted.

Task 2 officially contributes twice as much as Task 1 — yet most candidates split their hour as if the tasks were equal. And the descriptor lines most guides still quote died in May 2023.

Task 1 officially counts half as much as Task 2 — yet one missing element quietly caps your Task Achievement score. The descriptors name it, and most guides bury it.

The prompt tells you who you are writing to — and that one line fixes your register before your first sentence. Get it wrong and the descriptor wording is waiting for you.

A complete Task 2 opinion essay, read back against the four official criteria — quoting the May-2023 descriptor sentences that model-essay pages never actually open.

Your Writing band is four separate scores averaged — and each one fails in a different way. Diagnose which criterion is costing you before you 'practise writing' for another month.

A discussion question asks for three things — view one, view two, your opinion — and most essays deliver two. One full essay, read line by line against the May 2023 descriptors.

A complete Task 1 report on a checkable data set, read against the real May-2023 descriptors — which sentence is the overview, and which two numbers were deliberately left out.

A complete semi-formal letter about a neighbour's tree — then a line-by-line read against the actual Task Achievement wording, showing where each bullet is answered and why the tone holds.

Three problems, one thin solution bolted on — that mismatch is how problem-solution essays fail. A full sample where every solution answers a named cause, read line by line against the May 2023 descriptors.

Even on the computer-delivered test, IELTS Speaking is a live interview with a human examiner — and it is recorded. What each of the three parts asks of you, and what the four criteria actually say.

IDP says examiners can tell when you've memorised an answer — and it may influence your band. Here's a real cue card, a real two-minute answer, and the exact sentences you couldn't reuse.

Part 1 asks about home, work and study — so predictable that people script answers word-for-word. IDP says examiners can tell, and it may cost you. Here's what to do instead.

Part 3 asks abstract questions you've never rehearsed — which is exactly where the official Fluency descriptors split band 6 from band 7. One question, answered twice, shows the gap.

Your Speaking band is four equally-weighted numbers, and each one fails differently. Diagnose which criterion is actually costing you — with a recorded self-check for each — instead of vaguely 'speaking more fluently'.

Memorised place descriptions fit every place — which is why examiners spot them. A full spoken-style Part 2 answer about one unmistakable bench, read against all four criteria.