How Many Questions Do You Need for 99 Percentile in CAT?

There is no single correct number of attempts for 99 percentile. CAT 2025's overall benchmark was roughly 85 marks, but that score can come from very different combinations of coverage and accuracy.

Guide

There is no single correct number of attempts for 99 percentile. CAT 2025's overall benchmark was roughly 85 marks, but that score can come from very different combinations of coverage and accuracy.

Four valid-looking paths to an ~88 score

AttemptsCorrectWrong*AccuracyNet score*
30300100%90
3531488.6%89
4032880%88
48341470.8%88

*This intentionally conservative example assumes every wrong answer is an MCQ carrying −1. Wrong TITA/non-MCQ responses do not attract the same penalty.

The table shows why attempt count is a poor target by itself. The student touching 48 questions does far more work than the student attempting 35, yet ends near the same net score because of leakage.

99 percentile needs breadth—but controlled breadth

At this level you normally need enough competence to find scoring opportunities across the paper. That does not mean solving every difficult question. In VARC, avoid turning two close options into a long debate. In DILR, the first set decision matters disproportionately. In QA, a fast first pass should collect questions whose approach is visible before you invest in longer algebra or geometry.

Build your own range from five mocks

Track: attempts, accuracy, easy questions missed, avoidable negatives and time spent on abandoned work. If accuracy falls below your normal band when attempts rise, the extra coverage is probably low quality. If accuracy remains high and you repeatedly leave approachable questions unseen, then increasing pace is justified.

For 99 percentile, your best attempt plan is the one that produces a repeatable 85–90+ capability with healthy sectionals, not the one with the biggest attempt number.

What separates a one-off 99 from a repeatable one

Look at the spread of your last five mock scores, not only the best. If the range is 65–90, you still have a stability problem even though one paper touched the target. Reduce that variance by protecting easy questions, avoiding low-confidence MCQ guesses and making the first DILR-set choice more deliberate. A smaller score spread is often the clearest sign that the 99-percentile level is becoming real rather than accidental.

Preparing for CAT?

See where you actually stand before changing your shortlist or study plan.

Why On Your Mocks?

A mock score tells you what happened. OYM helps you understand what went wrong and what to work on next.

Understand where marks were lost
Identify what needs revision
Know what to practise next
Move from score to improvement
Explore Test Series →

Related resources