CAT percentile predictor
Type your score and pick a year. You get an estimated percentile from the published score bands for that year, with the bands either side shown so you can see how wide the estimate is.
How the estimate works
Coaching institutes publish the score that matched each percentile after the results came out. We take the two published bands either side of your score and draw a straight line between them. The real curve is steeper near the top, so treat the number as a guide and read the bands next to it.
Year-by-year tables: CAT 2025 score vs percentile, CAT 2024 score vs percentile, CAT 2023 score vs percentile.
Why your real percentile can differ
- CAT runs in three slots. The IIMs scale each slot's scores so a harder paper is not a penalty. The scaled score, not your raw score, sets your percentile, and the formula is not published.
- The bands come from coaching analyses of results, not from an IIM table. Two institutes' 2025 figures differ by up to a mark.
- A mock score is not a CAT score. Mocks from different providers run easier or harder than the real paper. Our free CAT mock follows the official exam pattern and marking.
Score bands are overall only. Section-wise bands are not published widely enough to estimate each section, so we do not.
Common questions
Is 40 a good score in CAT?
Against CAT 2025 bands, 40 marks sits between the 80 percentile (38 marks) and the 85 percentile (44.2 marks), so about the 82nd percentile. In CAT 2024 the same score was about the 75th. Use the tool to check any score.
What score do I need for the 99 percentile?
About 85 marks in CAT 2025 and about 95 marks in CAT 2024, according to published bands. It moves every year with paper difficulty.
How is the CAT percentile calculated?
Your raw section scores are scaled across slots, added up, and ranked against every candidate. Your percentile is roughly the share of candidates you scored above. The IIMs do not publish the exact scaling formula.
Which CAT percentile predictor is most accurate?
None can be exact, because the scaling formula is not public. A predictor is only as good as the band data behind it. We show the source and the bands either side so you can judge the estimate yourself.
Take a free mock and get a score to check
Sources
Last checked 8 Nov 2026. Found an error? Tell us and it goes on the errata page.