- A CAT percentile is a position among all test-takers, not a fixed mark, so the same raw score can end at different percentiles.
- CAT 2026 is held in three sessions on 29 November 2026 with different papers, so raw scores have to be put on one scale before percentiles are worked out.
- Marking is +3 for a correct answer and -1 for a wrong MCQ; non-MCQ (TITA) answers have no penalty and skipped questions score zero.
- Any table of 'marks for 99 percentile' is an estimate. Read mock results as a range and trust only your own scorecard.
CAT marks vs percentile is not a fixed table. A raw score becomes a percentile only after it is compared with everyone else who took the test, and CAT 2026 is held in three sessions on one day, 29 November 2026, with a different paper in each. That is why one score can land at two percentiles, and why every 'marks for 99 percentile' list is a guess.
This post explains the three steps from raw score to percentile, what the +3 and -1 marking does to the raw score you can reach, and how to read a mark range without being misled. It gives no mark target for 99 percentile, because no official figure exists for CAT 2026. Checked on 6 November 2026.
Why CAT marks vs percentile can differ for the same raw score
A percentile is a position, not a mark. It tells you what share of test-takers scored at or below you. If the group is stronger or the paper is harder, the same marks sit at a different position.
CAT 2026 is a computer-based test held in three sessions on one day, and each session has its own questions. Comparing raw marks directly would punish candidates who got a harder paper, so scores from different sessions have to be put on one scale before anyone is ranked. The official bulletin we checked does not describe the method, so read the details below as how earlier cycles worked and confirm them on your 2026 scorecard.
Raw score, scaled score, percentile: the three steps
- Raw score. Marks from your own answers, before any adjustment. For MCQs it is 3 × correct - 1 × wrong, plus 3 for each correct non-MCQ answer.
- Scaled score. The raw score adjusted so that results from different sessions can be compared. Earlier CAT scorecards showed a scaled score next to the percentile. The CAT 2026 bulletin does not spell out the method.
- Percentile. The share of all test-takers whose scaled score is at or below yours. A rough conversion is (N - rank) / N × 100, where N is the number of test-takers. The exam body decides how ties are treated.
- (100000 - 2000) / 100000 × 100 = 98.
- One percentile point is 1% of 100000 = 1,000 candidates.
Answer: About the 98th percentile. A few marks can move you across a point that holds about 1,000 candidates.
Colleges look at percentiles overall and by section. IIM Ahmedabad's CAT 2026 shortlisting policy, for example, sets minimums for both.
What the +3 and -1 marking does to the raw score you can reach
| Answer | Raw marks |
|---|---|
| Correct MCQ | +3 |
| Wrong MCQ | -1 |
| Correct non-MCQ (TITA) | +3 |
| Wrong non-MCQ (TITA) | 0 |
| Not attempted | 0 |
Accuracy pays twice: a right answer adds 3 and avoids the -1. A candidate with 40 correct and 6 wrong MCQs scores 3 × 40 - 6 = 114. A candidate with 38 correct and none wrong scores 3 × 38 = 114. The raw score is the same, the profiles are not, and both then face the same session adjustment.
Why every 'marks for 99 percentile' table is an estimate
- The conversion depends on the paper's difficulty and on how strong that year's test-takers are. Both change every cycle.
- Published tables are coaching or expert estimates. They are not official results.
- A table in raw marks cannot show the session adjustment, which sits between raw marks and percentile.
- Official percentile-to-score data for CAT 2026 will not exist until results are out, which the bulletin says is likely in the first week of January 2027.
Tables for the same percentile often disagree with each other, sometimes on the same page. That is a sign of guessing at a range, not of a precise rule.
Read a range, not a single number
| Source of the number | What it can tell you | What it cannot tell you |
|---|---|---|
| Coaching 'expected marks' table | A rough size of the marks involved, for the year it was written | Your session, or this year's difficulty |
| A mock provider's percentile | Your position among that provider's users | Your position among all CAT 2026 test-takers |
| Your CAT scorecard (after results) | Your actual percentile | How a given college will weigh it |
- Average = (93 + 95 + 92 + 96 + 94) / 5 = 470 / 5 = 94.
- Range = 96 - 92 = 4 points.
Answer: A band of 92 to 96, centred on 94. Plan around the band, not the best score.
What to do if you think you got the harder slot
You cannot know for certain which session was harder, and candidates in every session tend to feel theirs was tough. Putting sessions on one scale is meant for exactly this case, so there is nothing to fix on your side until the scorecard arrives.
- Do not compare raw scores with a friend who sat another session.
- Do not change your college list on a feeling about difficulty. Wait for the scorecard.
- Spend the wait on the other exams still open. XAT 2027 is on 3 January 2027, and its pattern, with negative marking of -0.25 and a Decision Making section, is different from CAT.
How to sanity-check your expected percentile from mocks
- Take at least three full mocks under CAT conditions: 120 minutes, three sections of 40 minutes, no switching between sections, +3 and -1 marking.
- Write down raw score, accuracy and percentile for each, and review each mock within 24 hours using the mock analysis guide.
- Take the lowest and highest percentile as your band, and the average as its centre.
- Shade the band down a little if the provider's test-takers are likely to be a weaker group than the real exam's. Nobody publishes that gap, so treat it as a margin, not a number.
The free CAT Compass mock uses CAT-style marking and gives an indicative percentile band from your net marks as a share of the maximum. It is shorter than the real CAT and lets you move between sections, so use it to track progress, not to predict your result.
What the official scorecard will and will not tell you
The bulletin says CAT 2026 scorecards will be accessible on the CAT website, and the score is valid until 31 December 2027. Each IIM sets its own eligibility, weights and selection process, so the same percentile is worth different amounts at different colleges.
- It will show your scores and percentiles as the CAT team reports them.
- It will not tell you which college you can get, or any college's cutoff.
- It will not settle how a college weighs section scores against overall scores.
For a first look at colleges by percentile band, see CAT percentile vs MBA colleges. Those lists are estimates for General category candidates, and the MBA Predictor narrows them using your category, profile and budget.