CAT 2026 Marking Scheme

CAT has multiple-choice questions and non-MCQs, often called TITA (Type In The Answer). Separate the two before counting wrong answers. Deducting one mark for every incorrect response will understate your score whenever you have a wrong TITA answer.

The official CAT FAQ does not fix the number of non-MCQs in advance. Use the question counts in the paper you actually attempt when checking maximum marks or unattempted questions.

How to Calculate Your CAT Score

Calculate each section first

For each of VARC, DILR and QA, add correct MCQs and correct TITA answers, multiply by three, then subtract only wrong MCQs. Adding these three sectional raw scores gives the overall raw score.

In this worked example, the entries are illustrative response counts, not a forecast of the 2026 question distribution.

VARC: (8 + 2) x 3 - 3 = 27. The overall total is (21 + 5) x 3 - 9 = 69 marks. Any wrong TITA answers contribute zero and do not change that total.

Keep marks, attempts and accuracy separate

Suppose these responses also include three wrong TITA answers. There are 38 attempts: 26 correct, nine wrong MCQs and three wrong TITA. Attempt accuracy is 26 / 38 x 100 = 68.42%. Accuracy alone cannot tell you the score because wrong MCQs and wrong TITA answers have different penalties.

Calculate Marks from Your Response Sheet

Use your recorded responses and the applicable answer key together. A solution you remember choosing is not a substitute for the answer saved in the response record. The CAT response sheet guide explains the download and score-checking process; the 2026 response sheet will follow the exam.

1. Match each response to the same question and session in the key. Use question identifiers where supplied, rather than relying only on display order.
2. Classify the response as correct MCQ, wrong MCQ, correct non-MCQ, wrong non-MCQ or blank.
3. Calculate each section separately. Check that correct plus wrong answers do not exceed that section's question count.
4. Add the sectional totals. If the applicable key is revised, recalculate the affected responses instead of adding a guessed adjustment.

Keep blanks separate from wrong answers. They are worth zero in both formats, but counting them as attempted questions would lower your reported accuracy.

Why Your Calculated Score May Change

A change from wrong to correct does not always add just three marks. For an MCQ, the earlier -1 is removed and +3 is awarded, creating a four-mark difference. For a TITA answer, the change is from zero to +3.

A dropped or disputed question is a different case. Apply the treatment announced for that question; do not assume every candidate automatically receives three marks. The calculator below applies ordinary correct, wrong and blank scoring, not a question-specific adjustment.

Raw Score, Scaled Score and Percentile

Your response counts are enough to calculate raw marks. They are not enough to reproduce score equating or to place you among every other CAT candidate. For example, 69 raw marks do not mean 69 percentile. Dividing 69 by a paper's maximum marks gives a marks percentage, not a percentile.

IIM Calcutta's MBA 2027-29 policy makes this distinction clear: it checks sectional and overall CAT percentiles first, then uses the total overall scaled score in its interview-shortlisting formula. See CAT score vs percentile for the difference between a score target and an admission cutoff.

Use Your Mock Score to Find the Mark Loss

Compare wrong MCQs separately from wrong TITA answers. Correcting one wrong MCQ in a later attempt improves the raw total by four marks; solving an unanswered question correctly adds three. A high attempt count with many wrong MCQs can therefore produce the same score as fewer, more accurate attempts.

These combinations all produce 75 marks. They are not 2026 percentile targets. For a section-level review, record the source of each error: an incorrect calculation, a misread condition, an unsupported RC inference or an unfinished set. This makes the next review about the responses that cost marks, rather than the attempt count alone.

1. Marking Scheme
2. Calculate Your Raw Score
3. Check Your Response Counts

1. When an Answer Changes
2. What Your Total Means