Celpip AIPractice Test Try a scored task

Free tool

CELPIP Score Calculator

Enter how many CELPIP Reading or Listening questions you answered correctly out of 38. You get an estimated CELPIP level — which is also your CLB level — shown as a band, with the published table it came from and the reason the band is not narrower.

Estimate your level

The two published tables, side by side

Two artefacts from Paragon carry a Reading conversion, and they do not agree. The sample test’s own code uses hard, non-overlapping cut-offs. The printed Reading Pro Study Pack uses wider bands that deliberately overlap by one or two raw points at every boundary. The overlap is the point: it is Paragon saying, in print, that the boundary is not sharp.

They cannot be reduced to each other. The Study Pack places a raw 24 at level 7; the code places the same 24 at level 6. Our calculator follows the code, because it is executable rather than typeset, and then widens by one level in the direction your score leans inside its band. That widening lands on exactly the raw scores where the Study Pack’s own bands overlap — 10–11, 15–16, 19–20, 24–25, 28, 31 and 33 — so we admit uncertainty in the same places Paragon does.

Reading – Paragon sample-test cut-offs
Raw correct / 38CELPIP level
34–3810–12
32–339
29–318
26–287
21–256
17–205
12–164
8–113
0–7M
Reading – Study Pack printed bands
CELPIP levelRaw correct / 38
10–1233–38
931–33
828–31
724–28
619–25
515–20
410–16
38–11
M–20–7

Left: Paragon Free Online Sample Test, client-side function calcReadScore. Right: CELPIP Reading Pro Study Pack (PDF, © 2022 Prometric).

Listening has only one published table

The sample test carries a separate set of Listening cut-offs, and they are not the Reading ones — a raw 30 is Reading level 8 but Listening level 7. We found no published Study Pack conversion table for Listening in any source we checked, so there is no second column here. An invented one would be worse than none.

Listening – Paragon sample-test cut-offs
Raw correct / 38CELPIP level
36–3810–12
34–359
32–338
29–317
24–286
19–235
13–184
8–123
0–7M

Second Listening table: not found. If Paragon publishes printed Listening bands the way it does for Reading, we have not located them. This space stays empty until we can cite one.

CELPIP level, CLB level and CEFR

For levels 3 to 12 the CELPIP level is the CLB level, so a raw-score estimate is a CLB estimate. CELPIP 2 falls into the CLB 1–2 band.

CELPIP level, CLB level and CEFR
CELPIPCLBCEFR
1212C2
1111C1
1010C1
99B2
88B2
77B2
66B1
55B1
44A2
33A2
21–2A1

Score chart for every raw score

The band tables above are the source. If you want the row-by-row chart for a particular section — every raw score from 38 down to 0, with the estimate next to it — those live on their own pages:

CELPIP Reading score chart   CELPIP Listening score chart

Questions people actually ask

Is this an official CELPIP score?

No. This is an independent estimate built from published material. Only Paragon Testing Enterprises issues official CELPIP scores, and the official raw-score conversion is not published.

Where do the thresholds come from?

From Paragon’s own free CELPIP-General sample test, which scores itself in your browser. We read the cut-offs out of its scoring functions: calcReadScore for Reading and findListScore for Listening. That is Paragon’s running code rather than anybody’s guess — but it is the sample test’s estimator, not the operational scoring engine.

Why does the calculator show a range instead of one level?

Because a raw total does not determine a level on its own. The real exam is equated across forms and weighted by question difficulty, and Paragon’s two public tables disagree with each other by a level at several points. A single number would claim more precision than anyone outside Paragon has.

Why do Reading and Listening use different thresholds?

They are different tables in the source, and they should be. The same raw total lands one level lower in Listening than in Reading — a raw 30 is Reading level 8 but Listening level 7. That matches the official 2024 distribution, where the Listening mean is 8.22 and the Reading mean is 7.21: candidates score higher on Listening, so a given level takes more correct answers.

What does the M band mean?

M is what Paragon’s code returns for a raw total of 0 to 7. It is a refusal to report a level rather than a level itself, so we show it as “below 3”.

Can I convert the score from a single part?

No. These tables convert a full 38-question section. There is no published conversion for an 11-question part on its own, so converting one would mean inventing a table.

Is a CELPIP level the same as a CLB level?

For levels 3 to 12 they line up one to one: CELPIP 9 is CLB 9. CELPIP 2 falls into the CLB 1–2 band. The full mapping, including CEFR, is on this page.

Where these numbers come from