Short answer: the Cambridge IELTS series is a set of past-paper collections, each containing four complete tests with answer keys.
They are the most format-accurate practice available, and they are measurement tools rather than courses — a volume tells you where you stand, and teaches you nothing on its own. Start with one of the recent volumes for current format accuracy, then work backwards through older ones.
We publish IELTS practice ourselves, so read the following knowing that: the recommendation below is to buy the official books, because for format fidelity nothing else comes close. What we would argue against is using only them, and the reason is arithmetic rather than sales.
What a Cambridge volume actually contains
Each numbered volume in the series holds four complete tests — Listening, Reading, Writing and Speaking — plus answer keys, audio, sample answers for the Writing tasks, and the tapescripts.
Academic and General Training editions are published separately, and the General Training volumes usually share the Listening sections with their Academic counterparts because Listening is common to both modules.
That is the whole product. There are no strategy chapters, no explanation of why a distractor worked, and no per-question-type breakdown of your score. Those omissions are deliberate — the books exist to reproduce the exam, not to comment on it.
Which volume, and in what order
| Situation | Use | Why |
|---|---|---|
| You have never sat a full paper | The most recent volume you can buy | Format, timing and question-type balance are closest to what you will actually sit. |
| You want maximum practice volume | Recent volume first, then work backwards | Older volumes remain valid English practice; only the format packaging has moved on. |
| You are sitting computer-delivered | Recent volumes, plus the official online familiarisation test | A paper book cannot reproduce the on-screen interface. Use the free official familiarisation for that. |
| You are on General Training | The General Training edition, not the Academic one | Reading differs in text type and in the conversion table, which is stricter for GT. |
| You have one week left | One volume, one full timed paper, then targeted drilling | Sitting four more papers in a week measures four times and improves nothing. |
On "which volume is hardest": there is no published difficulty ranking, and candidates report different volumes as hardest depending on which topics happen to suit them. Anyone offering a definitive ranking is offering an opinion.
What does vary measurably is format currency — the newest volumes reflect the current test, older ones reflect the test as it was.
How to use a book so it teaches
Most candidates use these volumes in a way that guarantees a plateau: sit a test, check the score, feel encouraged or discouraged, sit the next one. Four tests later they have four data points and no new skill.
- Sit the paper under real conditions. Full timing, no pausing the audio, no dictionary. A relaxed score is not information.
- Mark it and convert the raw score honestly. No credit for "I nearly had that one". Use the conversion for your module — the Reading calculator carries both the Academic and General Training tables.
- Sort every wrong answer by question type, not by passage. This is the step that turns a measurement into a diagnosis. Three losses in matching headings and one in True/False/Not Given is a completely different problem from the reverse.
- For each wrong answer, name the trap. Was it a synonym you accepted too readily, an overstated claim, a distractor placed before the answer, a word limit exceeded? An error you cannot name will repeat.
- Drill that one type until the method is automatic — not by sitting another whole paper, which mixes eleven types and hides which one is failing.
- Only then sit the next test, and compare by question type rather than by total.
Step 3 is where the books stop helping, because they give you a key rather than an analysis. It is also the step that actually moves a band.
The arithmetic problem
Four tests per volume. If you are drilling a weak question type, one volume gives you perhaps four or five instances of it — nowhere near enough repetitions to make a method automatic, and once you have seen a passage you can never sit it cold again.
This is the real limitation of book-only preparation, and it is not a criticism of the books. They were never designed to supply unlimited repetitions of one question type; they were designed to reproduce whole exams faithfully, which they do better than anything else.
The sensible split is therefore: official books for measurement, unlimited material for drilling. Sit a Cambridge paper every two or three weeks to check the band, and spend the time between on the specific types that lost you marks.
Where to buy, and what to avoid
Buy from Cambridge University Press & Assessment or any legitimate bookseller. The audio comes with the book or through the publisher's own download.
Avoid the PDF sites. Scanned Cambridge volumes circulating as free downloads are straightforwardly copyright infringement, the audio is usually missing or re-encoded badly, page scans lose the layout that makes the timing realistic, and the sites hosting them are a security risk.
We do not link to them and will not. If cost is the constraint, the free official familiarisation test plus free legitimate practice is a better answer than a bad scan.
To be explicit: we are not affiliated with Cambridge University Press & Assessment, IDP or the British Council, and nothing here is endorsed by them.
When you finish the series
You will, eventually — that is what a finite set of books means.
What to do when you have finished all the Cambridge tests covers that in full: the short version is that you need unlimited fresh material for the two or three question types still costing you marks, and one more measured paper before test day.
Our free reading practice tests are full 40-question papers with a key that names the trap behind each wrong option, and practice by question type generates fresh passages so the repetitions do not run out.