DET Mock Test 3
The third disjoint set, with its own passages, dictations, conversations and prompts. Same difficulty range and the same scoring, so all three scores are directly comparable.
- About an hour, eight sections, run in the real order
- All 11 question types — nothing here appears in the other 2 tests
- Eight subscores and per-answer feedback the moment you finish
Mock Test 3 is a full sitting, not a sample. It opens with the device check the real test opens with, runs the eight sections in the order Duolingo runs them, and ends with a scored report. Budget an uninterrupted hour and a working microphone.
Free for a limited time. No card, and no account needed to take it or to see your score.
What this test covers
Eight sections, in the order you will meet them on the day. The number of questions in the adaptive sections varies with how you answer — that is the engine narrowing in on your level, and it is why two people rarely see the same number of items.
| Section | Task types | Roughly |
|---|---|---|
| 1 · Read & Select | Read & Select | 15–18 words, 5 sec each |
| 2 · Fill in the Blanks | Fill in the Blanks | 6–9 sentences |
| 3 · Mixed Language Skills | Listen & Type, Read & Complete | 9–15 items, 24 min total |
| 4 · Interactive Reading | Interactive Reading | 2 passages, 5 steps each |
| 5 · Interactive Listening | Interactive Listening | 2 conversations |
| 6 · Writing | Write About the Photo, Writing Sample | 4 written responses |
| 7 · Speaking | Speak About the Photo, Interactive Speaking | Photo + spoken conversation |
| 8 · Writing & Speaking Sample | Writing Sample, Speaking Sample | 8 min total |
What you get at the end
A predicted overall
On the real 10–160 scale, with a confidence range, because a single sitting cannot place you to the point.
All eight subscores
The four skills and the four integrated scores Duolingo reports, each computed from the items that actually measured it.
Every answer, reviewed
Question by question, with the model answer beside yours and the one change that would have lifted it.
Your writing and speaking graded
Rubric-scored on task fulfilment, grammar, vocabulary and coherence, with the evidence quoted back to you.
Your recordings kept
Play back what you actually said. Stored on your device, never uploaded.
Your three biggest leaks
Ranked by how much fixing each one moves your overall, with a drill for each.
How to get a number worth trusting
Give it a clean hour
The score assumes an uninterrupted sitting. Stopping halfway and resuming tomorrow measures something else.
Use a microphone
A quarter of your overall comes from speaking. Without a working mic those tasks cannot be graded and the estimate loses a whole skill.
Do not look anything up
The engine places you from what you can do unaided. A score you helped is a number you cannot use.
Finish every task
Partial credit is real and a blank is a certain zero. A rough answer beats no answer on every single task.
How Mock Test 3 differs
It does not differ in difficulty, and that is deliberate. A test that started harder would not measure a stronger candidate better — the engine already adjusts to you after every answer, so difficulty is set by your responses rather than by which test you picked. What changes between the 3 is the content: every passage, dictation, conversation, photograph and writing prompt in this test appears in no other.
That is what makes a second sitting worth taking. If the tests shared items, your score would climb because you remembered them, and you would have no way to tell that apart from genuine progress. Because they do not overlap, the difference between your first score and your third is information.
| Test 1 | Test 2 | Test 3 | |
|---|---|---|---|
| Questions shared with the others | None | None | None |
| Difficulty range | Full, A1 to C2 | Full, A1 to C2 | Full, A1 to C2 |
| Sections | All eight | All eight | All eight |
| Scored the same way | Yes | Yes | Yes |
The three are equated: simulated across a wide range of abilities, the same candidate scores within about two points on each — smaller than the variation you would see re-sitting the same test twice. So a 105 here means what a 105 means on either of the others.
What this score is and is not
This is a practice estimate, not an official result. It carries no standing with any institution and it cannot be sent anywhere. What it is good for is finding your weakest skill, telling you whether it is moving, and showing you whether you can hold concentration for a full hour — which is a real part of what the day tests.
Writing and speaking are graded by a model against a rubric. That is genuinely useful for spotting a pattern across several answers, and it is less reliable than a human rater on any single one. Treat one report as weak evidence and a consistent run of them as strong evidence. Where our engine is uncertain, it widens the range it reports rather than hiding the doubt behind a single confident number.
The official test costs $70 USD and results arrive in 2 days. Practising here is free for a limited time; sitting the real thing is not, which is the whole reason to find your weak spots before you book it.
Start a practice test
An hour under real conditions, all eight subscores, and feedback on every written and spoken answer.
Take a test