CAT DILR: Score Needed for 90 Percentile

For a 90 percentile target in DILR, CAT 2025 gives a practical reference of roughly 17 marks, with a sensible preparation band of 17–19. Treat that as a benchmark for training, not a promised CAT 2026 conversion.

Across recent CATs, the requirement has moved: about 17 in 2023, 22 in 2024 and 17 in 2025 under the research taxonomy. That volatility is normal. Section difficulty changes more sharply than students often expect.

Translate the score into questions

At perfect accuracy, the target is worth roughly 6 clean correct questions. Real attempts will normally be higher because you will not operate at 100% accuracy, and the mix of MCQ and TITA matters. The right takeaway is not “attempt exactly 6.” It is that a relatively small number of high-quality decisions can create a strong sectional percentile.

Roughly 6 clean correct answers' worth of score; one solid set plus additional correct questions can move the percentile sharply.

How to build this score

DILR is usually decided by set selection before it is decided by calculation. Spend the first few minutes comparing entry points, data clarity and the number of usable constraints before committing.

Run a simple two-pass review after every sectional. First, identify questions you should have scored: easy/medium questions that were missed, misread or reached too late. Second, identify questions you should not have spent time on. Improving both lists at the same time usually adds marks faster than learning a large new topic list.

When the target is not yet realistic

If your current DILR score is more than 8–10 marks below this band, spend the next two weeks on process rather than percentile. Improve one measurable thing: passage option elimination in VARC, set selection in DILR, or first-pass question selection in QA. If you are already within 4–6 marks, focus on repeatability and avoid turning a good section into an over-attempted one.

Your review should record not only whether you solved a set, but whether you selected it early enough. A correct set solved after wasting ten minutes elsewhere is still a selection problem.

Use the target as a guardrail for your overall plan. You do not need the same percentile in all three sections, but you do need enough balance that a single section does not close doors your total score would otherwise open.