How the GATE score is calculated
Applies to any GATE year. Source: GATE 2027 Information Brochure.
The brochure prints a formula, not a number — and a 350-to-900 scale whose two anchor marks are computed only after the examination is over.
IAE Academy
Resource Centre
A GATE score is not the marks you scored. It is those marks placed on a scale where 350 is pinned to the qualifying marks for your paper, and 900 is pinned to the mean marks of the top 0.1% or top 10 (whichever is larger) of the candidates who appeared in that same paper. Both of those anchors are computed after the examination, out of the marks of the people who sat it. That is why a past year's qualifying marks are not a rule you can plan against: they were that year's arithmetic on that year's candidates. The rule is the formula, and this post works through it.
Source: GATE 2027 Information Brochure, p. 28, as on 15 September 2026
The qualifying mark is a formula, not last year's number
The brochure prints it in one line: qualifying marks for the GENERAL category = max(25, min(40, μ + σ)), where μ is the mean and σ is the standard deviation of the marks of all the candidates who appeared for that test paper. Read it from the inside out. Take the mean of everyone's marks in your paper, add one standard deviation, and you have a provisional figure. If that figure is above 40, the cap pulls it down to 40. If it is below 25, the floor lifts it to 25. Between the two, it stands as computed. Three consequences follow. The number is worked out separately for each of the 30 test papers. Its only inputs are the marks of the candidates who sat that paper, so no earlier year's figure enters it. And it can never land outside 25 to 40 marks, out of the 100 marks a test paper carries.
Sources: GATE 2027 Information Brochure, p. 28, as on 15 September 2026; GATE 2027 Information Brochure, p. 23, as on 15 September 2026; GATE 2027 Information Brochure, p. 2, as on 15 September 2026
What the floor and the cap do
| Illustrative paper | Mean μ | SD σ | μ + σ | Qualifying mark, General | Which part of the rule decided it |
|---|---|---|---|---|---|
| A, a hard paper | 14.0 | 8.0 | 22.0 | 25.0 | The floor. max(25, …) lifts 22.0 to 25.0 |
| B, a middling paper | 22.0 | 12.0 | 34.0 | 34.0 | Neither. 34.0 sits inside 25 to 40, so μ + σ stands |
| C, an easy paper | 32.0 | 15.0 | 47.0 | 40.0 | The cap. min(40, …) pulls 47.0 down to 40.0 |
Source: GATE 2027 Information Brochure, p. 28. The μ and σ values are IAE Academy's illustrative inputs, chosen only to show how the floor and the cap behave. They are not GATE data, not any paper's real statistics, and not a forecast.
The other categories are fixed fractions of the General figure
The brochure calls these the standard ratio. Qualifying marks for OBC-NCL and EWS are nine-tenths of the General figure; for SC, ST and PwD they are two-thirds of it. So the numbers move together: settle the General figure for a paper and the rest follow. On illustrative paper B above, whose General figure came to 34.0, that would be 30.6 for OBC-NCL and EWS and 22.67 for SC, ST and PwD. That is IAE Academy's arithmetic on an illustrative input, not a GATE figure. One detail is easy to miss. The score formula below uses Mq, which the brochure defines as the qualifying marks for the General Category candidate in the paper. The category ratios decide who qualifies; the score scale is anchored at the same point for everybody.
Source: GATE 2027 Information Brochure, p. 28, as on 15 September 2026
The score formula
GATE Score = Sq + (St − Sq) × (M − Mq) / (M̄t − Mq). Sq is fixed at 350 and St is fixed at 900, so the span (St − Sq) is always 900 − 350 = 550 score points. The fraction is simply how far your marks travelled from the qualifying mark towards the top-0.1% mean: it is 0 at the qualifying mark and 1 at the top-0.1% mean. Multiply that fraction by 550, add 350, and you have the score. Of the five symbols, one is about you.
Source: GATE 2027 Information Brochure, p. 28, as on 15 September 2026
The five symbols
| Symbol | What the brochure defines it as | Where the number comes from |
|---|---|---|
| M | The marks obtained by the candidate: actual marks for single-session papers, normalised marks for multi-session papers | You, in that paper |
| Mq | The qualifying marks for the General Category candidate in the paper | Computed after the exam, as max(25, min(40, μ + σ)) |
| M̄t | The mean of marks of the top 0.1% or top 10 (whichever is larger) of the candidates who appeared in the paper; for multi-session papers, taken across all sessions. The bar over the M means it is an average | Computed after the exam, from the top candidates' marks |
| Sq | The score assigned to Mq | Fixed in the brochure at 350 |
| St | The score assigned to M̄t | Fixed in the brochure at 900 |
Source: The middle column is the GATE 2027 Information Brochure's own wording, shortened, from section 8.1, printed p. 28, read on 15 September 2026. The last column, and the note that the bar denotes an average, are IAE Academy's gloss on where each number comes from; the brochure prints the definitions only.
A worked example, start to finish
Take illustrative paper B, whose qualifying mark came out at 34.0, and suppose the mean marks of its top 0.1% come to 84.0. The gap between the two anchors is 50.0 marks, and that gap is worth 550 score points. So in this one illustration a mark is worth 11 score points, and the whole calculation collapses to: score = 350 + 11 × (your marks − 34.0). A candidate who scores the qualifying mark gets 350, because the fraction is zero. That is what 350 means. It is not a poor score; it is the definition of the qualifying line, and it is the floor of the scale by construction.
Source: GATE 2027 Information Brochure, p. 28, as on 15 September 2026
Marks to score, on one illustrative paper
Illustrative paper B: qualifying mark Mq = 34.0, top-0.1% mean M̄t = 84.0, so the span is 50.0 marks
| Marks in the paper (M) | (M − Mq) ÷ (M̄t − Mq) | 350 + 550 × that = GATE score |
|---|---|---|
| 30.0, below the qualifying mark | Negative, since M is below Mq | The brochure states that scorecards are available for download by the qualified candidates |
| 34.0, the qualifying mark | 0 ÷ 50 = 0 | 350 |
| 40.0 | 6 ÷ 50 = 0.12 | 416 |
| 50.0 | 16 ÷ 50 = 0.32 | 526 |
| 60.0 | 26 ÷ 50 = 0.52 | 636 |
| 84.0, the top-0.1% mean | 50 ÷ 50 = 1 | 900 |
Source: Formula and the 350 and 900 anchors from the GATE 2027 Information Brochure, printed p. 28, read on 15 September 2026, as is the statement that scorecards are available for download by qualified candidates. The qualifying mark of 34.0 and the top-0.1% mean of 84.0 are IAE Academy's illustrative inputs, and the score column is IAE Academy's arithmetic on them. No GATE 2027 figures of either kind exist yet.
These numbers are arithmetic, not results
Every μ, σ, qualifying mark and top-0.1% mean in the tables above is an input IAE Academy chose to illustrate the rule, and every score worked out from them is IAE Academy's arithmetic. Only the formula itself, the 350 and 900 anchors and the 25 and 40 bounds are the brochure's. None of the figures IAE Academy supplied is an official figure, a predicted cut-off or an expected score. Nobody can know GATE 2027's qualifying marks in advance, because they are computed from the mean and the standard deviation of the marks of everyone who sits that paper, and those two statistics do not exist until the examination is over and the answers have been evaluated. Any page publishing a GATE 2027 cut-off today is publishing a guess.
Why the same marks give different scores in different papers
The conversion rate in the worked example, 11 score points per mark, came out of one specific gap: 84.0 minus 34.0. Change either anchor and it changes. Had the same qualifying mark of 34.0 sat against a top-0.1% mean of 70.0, the gap would be 36.0 marks and a mark would be worth about 15.3 score points, which is 550 divided by 36 and again IAE Academy's arithmetic. A paper in which the strongest candidates finish closer to the qualifying line packs the same 550 points into fewer marks. This is the concrete meaning of the brochure's own description of the score as reflecting the relative performance level of the candidate in a particular test paper, and it is why a score carried across two different papers says less than it appears to.
Sources: GATE 2027 Information Brochure, p. 5, as on 15 September 2026; GATE 2027 Information Brochure, p. 28, as on 15 September 2026
Normalisation applies only to multi-session papers
Some test papers may be held in more than one session. Where that happens, a candidate's raw marks are first converted into normalised marks, and it is the normalised mark that enters the score formula as M. Where a paper runs in a single session there is no normalisation at all, and the raw marks go straight in. The brochure gives the conversion as a formula; in words it works like this. Each session has two landmarks of its own: the sum of that session's mean and standard deviation of marks, and the average marks of the top 0.1% in that session. The paper as a whole has the same two landmarks, computed across every session. Normalisation maps each session's pair onto the all-sessions pair, and every mark in that session is stretched or compressed linearly between them. Read the formula and the direction follows: where a session's own two landmarks both sit below the all-sessions pair, the marks between them come out higher after normalisation, and where both sit above, lower. That reading is IAE Academy's; the brochure prints the formula and not its consequences.
Source: GATE 2027 Information Brochure, p. 27, as on 15 September 2026
The assumption normalisation rests on, in the brochure's own words
The brochure states its assumption rather than burying it: normalisation "is done based on the assumption that in multi-session GATE papers, the distribution of abilities of candidates is the same across the sessions". It justifies that on three grounds — the number of candidates appearing in multi-session papers will be large, the procedure for allocating candidates to sessions is random, and the number of candidates allotted to each session is comparable. Note what this method does not do: it does not grade individual questions for difficulty. It assumes the candidates in the different sessions were equally able as groups, and treats the remaining difference in their marks as a difference between the question papers.
Source: GATE 2027 Information Brochure, p. 27, as on 15 September 2026
What to do with this while you are still preparing
- Set targets in marks, not in score. Every test paper has 65 questions carrying 100 marks, of which 10 questions carrying 15 marks are General Aptitude. A score target cannot become a marks target until Mq and M̄t exist, and they exist only after the results.
- Count marks, not attempts. A wrong 1-mark MCQ costs one-third of a mark and a wrong 2-mark MCQ two-thirds; MSQ and NAT questions carry no negative marking; no question carries partial marks. M in the formula is what survives that.
- Treat qualifying and scoring well as two targets. The qualifying mark lands between 25 and 40 out of 100 by construction, while the 900 anchor sits wherever the strongest candidates in your paper finish.
- Do not plan around a predicted cut-off. Until the marks of everyone who sat your paper exist, μ, σ and the top-0.1% mean do not exist either, and all three are inputs to the number being predicted.
- Check whether your paper ran in multiple sessions before reading much into a raw mark. The brochure says only that examination for some test papers may be held in multiple sessions; it does not name which ones.
Sources: GATE 2027 Information Brochure, p. 23, as on 15 September 2026; GATE 2027 Information Brochure, p. 4, as on 15 September 2026; GATE 2027 Information Brochure, p. 2, as on 15 September 2026
Check it against the brochure yourself
The score and normalisation rules above are section 8 of the GATE 2027 Information Brochure, printed pages 27 and 28, published by IIT Madras as the organising institute on the official GATE 2027 website; the pattern and marking figures quoted earlier in this post are printed pages 4 and 23 of the same document. It is short: the score and normalisation rules take under two pages, and the qualifying-marks rule is four lines. When a page quotes you a GATE 2027 cut-off, open those two pages and ask which symbol in them the number is meant to be. Application dates, fees, eligibility and how long a score stays valid are a separate matter, set out in our GATE 2027 application checklist at /resources/gate-2027-application-checklist.
Sources: GATE 2027 Information Brochure, p. 27, as on 15 September 2026; GATE 2027 Information Brochure, p. 28, as on 15 September 2026; GATE 2027 Information Brochure, p. 4, as on 15 September 2026; GATE 2027 Information Brochure, p. 23, as on 15 September 2026
From the Resource Centre
GATE qualifying marks, 2024–2026
The qualifying marks for every GATE paper in 2024, 2025 and 2026, by category, from the organising institutes' own Statistical and Performance Reports, with how many registered, appeared and qualified, and what a qualifying mark does and does not get you.
ANY GATE YEARGATE 2027 application checklist
Every date, fee and eligibility rule for the GATE 2027 application, from the official GATE portal and brochure, as on 5 September 2026.
GATE 2027GATE 2027 Official Syllabus — All Papers
The official GATE 2027 syllabus, all 30 papers. Know the exact topics, sections and weightage before you plan a single day of prep. The new Robotics and Automation (RA) paper is not in this set yet.
GATE 2027GATE CS previous-year papers 2018–2026
Nine years of GATE Computer Science & IT question papers (2018–2026), as the organising institutes published them — including both slots for recent years. Learn the real pattern and build exam-day speed. The 2018, 2020, 2022 and 2023 papers come with the organising institute’s own answer key.
ANY GATE YEARNext step
Common questions
What are the qualifying marks for GATE 2027?+
How is the GATE score calculated from marks?+
Is the GATE score out of 1000?+
Does normalisation apply to my GATE paper?+
Why do two candidates with the same marks in different papers get different GATE scores?+
Can anyone predict the GATE 2027 cut-off in advance?+
IAE Academy
Resource Centre