GATE CS weightage: marks per subject in every official paper, 2014 to 2026

Pick a subject to see its marks in each of the 23 papers. The black line is its average. Notice how far single papers sit from that line.

S1 and S2 are the two sessions held in that year. Darker cells mean more marks; grey rows are fixed or combined blocks. Each column adds up to 100.

Marks:0 to 34 to 67 to 910 to 1213 or more

The core subjects are close together

Since 2021, all nine core subjects averaged between 6 and 10 marks. The largest and the smallest are about three and a half marks apart, which is two questions.

Single papers swing a lot

Algorithms carried 3 marks in 2022 and 10 in both 2026 sessions. Discrete Mathematics went from 4 marks in 2021 Session 2 to 15 in 2022, the biggest jump between consecutive papers. Even the two sessions of one year differ: in 2021, Discrete Mathematics had 9 marks in Session 1 and 4 in Session 2.

Some subjects have grown

Compared with 2014 to 2020, Computer Organization, Compiler Design and Digital Logic each gained about two marks on average, and Programming & Data Structures lost nearly three (12.1 down to 9.4). Calculus has carried only 1 to 3 marks since 2021.

The GATE 2027 CS syllabus in the official information brochure from IIT Madras trims Computer Networks and asks for more design in Digital Logic and Computer Organization, for example ALU and control-unit design, memory interfacing and cache mapping. Removed from Networks are, among others, ARP, DHCP and ICMP, UDP, and SMTP and FTP. The full list is on our GATE 2027 page.

No one can say yet how many marks the new topics will get, because no paper has been set on this syllabus. Questions from older papers on removed topics are still useful practice for the parts that remain.

GATE 2026 CS Session 1, Q22 (Algorithms, 1 mark). A hash table P[0..10] uses open addressing with linear probing and h(x) = (x + 7) mod 11. The keys 1, 13, 22, 15, 11, 24 are inserted in that order. Which of the positions 0, 10, 2 and 1 are empty afterwards?

The first four keys land in their home slots: 1 goes to 8, 13 to 9, 22 to 7 and 15 to 0. Key 11 hashes to 7, finds 7, 8 and 9 taken, and goes to 10. Key 24 hashes to 9, finds 9, 10 and 0 taken, and goes to 1. Only position 2 is still empty, which matches the official key. It was a multiple-select question, so a wrong answer would have cost nothing. See the whole paper

Every question in the 23 papers was tagged to one subject of the official CS syllabus and counted at its official marks (1 or 2). General Aptitude is counted as one block; Engineering Mathematics is shown as a total and by part. The source is the official question paper and answer key for each session. In these papers, no question was awarded to all candidates in the official keys.

Tagging takes judgement at the edges. The hashing question above is counted under Algorithms; a book that teaches hashing with data structures would count it under Programming & Data Structures. Borderline questions like this can move a subject's total by a few marks in a single paper, so read a difference of one or two marks between subjects with care. These counts are Lemyte's; the organising IIT publishes no subject-wise weightage.

Cover every subject: since 2021, no core subject has been worth fewer than 3 marks in a paper, and the “small” mathematics parts add up to 12 to 21 marks. Then let your own results set the order. A subject where you lose marks in timed practice matters more than its average.