GATE CS Previous Year Papers and How to Analyse Them

Previous year papers (PYQs) are the best guide to what GATE asks and how it asks it. This page tells you where to get the official papers and answer keys, and how to turn them into a topic-wise analysis that shapes your study plan.

Where to get official papers

• The current GATE website. GATE 2027 is organised by IIT Madras: gate2027.iitm.ac.in. Each year's official site has a section for previous question papers and answer keys.
• The GATE office of each IIT and IISc. The zonal GATE offices also host papers and keys of the years they organised.
• Your own response sheet. After the exam, candidates can view their responses and the provisional answer key on the GOAPS portal, and the final key is published with the results.

Avoid papers from unofficial sources that retype questions: they often contain typing errors, missing figures and wrong keys.

A topic-wise analysis in five steps

Make a table with one row per question and these columns: year, question number, marks (1 or 2), type (MCQ, MSQ or NAT), subject, topic, and whether you solved it.

1. Tag every question. Use the section names from the syllabus, and one topic name from within the section (for example Operating System → Paging).
2. Total the marks per subject for each year. The weightage moves a little from year to year, so look at the average over several years, not one paper.
3. Find the repeated topics. Some topics come almost every year (for example asymptotic complexity, CPU scheduling, normal forms, subnetting, cache mapping, DFA minimization). Make sure these are your strongest.
4. Mark your score per subject. Solve each paper in three hours under exam conditions, then record the marks you lost per subject.
5. Plan from the gaps. Spend the next week's study time where weightage is high and your score is low.

The program below shows how to total such a table. The data here is a made-up sample, not real GATE data: replace it with your own tags.

from collections import defaultdict

# (subject, marks, solved?) for each question you tagged. Sample data only.
tags = [
    ("Algorithms", 2, True), ("Algorithms", 1, False), ("OS", 2, True),
    ("OS", 2, False), ("CN", 1, True), ("CN", 2, True), ("TOC", 2, False),
    ("DBMS", 1, True), ("DBMS", 2, True), ("COA", 2, False),
]
total = defaultdict(int)
lost = defaultdict(int)
for subject, marks, solved in tags:
    total[subject] += marks
    if not solved:
        lost[subject] += marks
for subject in sorted(total, key=lambda s: -lost[s]):
    print(f"{subject:10} marks {total[subject]}  lost {lost[subject]}")

OS         marks 4  lost 2
TOC        marks 2  lost 2
COA        marks 2  lost 2
Algorithms marks 3  lost 1
CN         marks 3  lost 0
DBMS       marks 3  lost 0

Using PYQs at each stage

Each subject track on this site ends with a GATE practice set of original questions on the same patterns: for example Operating Systems, Computer Networks and Algorithms.

Practice