GATE · Exam basics
GATE score calculation: formula, normalisation and qualifying marks
On this page
GATE score calculation uses one formula: Score = 350 + 550 × (M − Mq) / (Mt − Mq). Here M is your marks, Mq is the General-category qualifying mark for your paper, and Mt is the mean marks of the top 0.1% of candidates, or of the top 10 candidates if that group is larger. Someone exactly at the qualifying mark scores 350, and someone at the top-group average scores 900, on a scale that runs up to 1000. If your paper ran in several sessions, your marks are normalised first, and only candidates at or above their category's qualifying mark receive a scorecard at all.
Raw marks, normalised marks and GATE score
These three numbers are easy to mix up. Each one is a step on the way to the next.
| Term | What it is | Who has one |
|---|---|---|
| Raw marks | Your marks out of 100 after negative marking | Everyone who appears |
| Normalised marks | Raw marks adjusted so sessions of different difficulty can be compared | Only papers that ran in more than one session |
| GATE score | A number up to 1000, calculated from raw or normalised marks | Only candidates who qualify |
Which papers run in more than one session is announced with the exam schedule, and the exam pattern page explains the rest of the format. In GATE 2026, CS and CE each had two sessions, published as CS1/CS2 and CE1/CE2. If your paper has a single session, there is no normalisation and your raw marks go straight into the score formula. The 2027 list of multi-session papers has not been announced yet.
The GATE score formula
The GATE 2027 brochure gives the score as Sq + (St − Sq) × (M − Mq) / (Mt − Mq), with Sq = 350 and St = 900. Because St − Sq = 550, that is the version at the top of this page.
| Symbol | Meaning |
|---|---|
| M | Your marks: raw for a single-session paper, normalised for a multi-session paper |
| Mq | Qualifying marks of the General category in your paper |
| Mt | Mean marks of the top 0.1% of candidates or the top 10 candidates, whichever group is larger, across all sessions |
| Sq, St | Fixed at 350 and 900 |
A worked example
These numbers are invented to show how the formula behaves. Suppose Mq = 30 and Mt = 80. Then Mt − Mq = 50, and each extra mark is worth 550 / 50 = 11 score points.
| Your marks (M) | Working | GATE score |
|---|---|---|
| 30 | 350 + 11 × 0 | 350 |
| 45 | 350 + 11 × 15 | 515 |
| 60 | 350 + 11 × 30 | 680 |
| 70 | 350 + 11 × 40 | 790 |
| 80 | 350 + 11 × 50 | 900 |
| 85 | 350 + 11 × 55 | 955 |
What this tells you:
- The same marks give different scores in different papers and years. If toppers in a paper score closer to the qualifying mark, the gap Mt − Mq is smaller and each mark is worth more points. A 650 in one paper is not the same achievement as a 650 in another.
- The top 0.1% mean sits at 900, not 1000. Candidates above that average score more than 900.
- A category's lower cut-off does not move the formula. As the brochure writes it, Mq is the General-category mark, so a candidate who qualifies on a lower category cut-off can end up with a score below 350.
- There is no separate General Aptitude or subject score. Both feed the one total, which is why the 15 marks of General Aptitude matter so much.
How normalisation works
When a paper is held in several sessions, a harder session would punish the candidates who drew it. GATE therefore rescales each session's marks. The brochure's formula, with the bars and hats dropped for readability, is:
Normalised marks = (Mtg − Mqg) / (Mti − Miq) × (Mij − Miq) + Mqg
| Symbol | Meaning |
|---|---|
| Mij | Raw marks of candidate j in session i |
| Mtg | Average of the top 0.1% of candidates across all sessions |
| Mqg | Mean plus standard deviation of marks across all sessions |
| Mti | Average of the top 0.1% of candidates in session i |
| Miq | Mean plus standard deviation of marks in session i |
The method assumes the spread of candidate ability is the same in every session. IIT Madras justifies this because candidates are allocated to sessions randomly and the sessions are of comparable size.
A worked example with two sessions
Again the numbers are invented. Across both sessions, Mtg = 80 and Mqg = 30. Session 1 turned out harder (Mti = 70, Miq = 26) and session 2 easier (Mti = 85, Miq = 33). Student A scored 50 raw in session 1, and Student B scored 60 raw in session 2. The score step uses Mq = 30 and Mt = 80, as above.
| Student A | Student B | |
|---|---|---|
| Session | 1 (harder) | 2 (easier) |
| Raw marks | 50 | 60 |
| Scaling factor | 50 / 44 = about 1.14 | 50 / 52 = about 0.96 |
| Normalised marks | (50 / 44) × (50 − 26) + 30 = 57.27 | (50 / 52) × (60 − 33) + 30 = 55.96 |
| GATE score | 350 + 11 × 27.27 = 650 | 350 + 11 × 25.96 = about 636 |
Student A has 10 fewer raw marks but a higher score, because the harder session's marks are stretched upward. This is the whole point of normalisation: your result does not depend on which session you were assigned. You also cannot choose a session, so there is nothing to gain from guessing which one will be easier.
Qualifying marks by category
The General qualifying mark for a paper is max(25, min(40, μ + σ)), where μ is the mean and σ the standard deviation of marks of all candidates who appeared in that paper. In plain words: the average plus one standard deviation, but never below 25 and never above 40 out of 100. Other categories get a fixed fraction of that.
| Category | Qualifying mark |
|---|---|
| General | max(25, min(40, μ + σ)) |
| OBC-NCL and EWS | 0.9 × General |
| SC, ST and PwD | 2/3 × General |
Three invented cases show how the floor and the cap work.
| Case | μ | σ | μ + σ | General | OBC-NCL / EWS | SC / ST / PwD |
|---|---|---|---|---|---|---|
| Low-scoring paper | 14 | 9 | 23 | 25 (floor) | 22.5 | 16.67 |
| Typical paper | 22 | 10 | 32 | 32 | 28.8 | 21.33 |
| High-scoring paper | 31 | 14 | 45 | 40 (cap) | 36.0 | 26.67 |
Older articles often say "25 or μ + σ, whichever is higher". That leaves out the cap of 40. The cap was already in effect for GATE 2026, where several papers' cut-offs sat at exactly 40 and none went above it.
GATE 2026 qualifying marks, for illustration
These are official GATE 2026 figures from the IIT Guwahati cut-off page. They change every year and say nothing about 2027.
| Paper | Appeared | General | OBC-NCL / EWS | SC / ST / PwD |
|---|---|---|---|---|
| CS | 2,11,020 | 30.0 | 27.0 | 20.0 |
| EC | 95,752 | 26.4 | 23.7 | 17.5 |
| CE | 76,385 | 28.7 | 25.8 | 19.1 |
| DA | 69,242 | 26.4 | 23.7 | 17.5 |
| EE | 65,801 | 27.7 | 24.9 | 18.4 |
| ME | 59,612 | 25.2 | 22.6 | 16.7 |
| IN | 12,424 | 39.8 | 35.8 | 26.5 |
| AR | 8,415 | 40.0 | 36.0 | 26.6 |
The CS row shows the multipliers cleanly: 0.9 × 30.0 = 27.0 and 2/3 × 30.0 = 20.0. Other rows differ slightly from the simple multiplication because the published figures are shown to one decimal place (for example, 2/3 of 40.0 is 26.67, shown as 26.6). In 2026, AR, EY, MT, PE and two XH sections sat exactly at 40, while AE, AG, BM, CH, CY and GE sat at the 25 floor.
Qualifying marks are not admission cut-offs. They decide only whether you get a scorecard.
From exam to scorecard
- Responses and answer key. Your responses are shown in the application portal, then provisional answer keys are published. You can contest a key during set dates by paying a fee. The 2027 brochure does not state the amount.
- Results on 19 March 2027 (the date is liable to change).
- Scorecard. Qualified candidates can download it free for a set window. After that it costs ₹500 per paper up to 31 December 2027, and no scorecards are issued from 1 January 2028. There are no hard copies, so download and save your PDF early.
- Validity. A GATE score is valid for 3 years from the result date.
What counts as a good GATE score
The qualifying mark is only the starting line. After that, "good" depends entirely on what you want to do.
- M.Tech, MS or PhD. Each admitting institute sets its own cut-offs by branch and category, and many add a written test or interview. Ministry of Education guidelines ask for at least 70% weightage to GATE in postgraduate admission, and the rest can be a test, interview or academic record.
- PSU recruitment. Each recruiter decides how to shortlist, with its own cut-offs, age limit and degree percentage, and often a group discussion or interview. Read the advertisement to see whether it uses GATE marks or GATE score.
The practical way to set a target is to look at the official qualifying marks and, for the institutes or companies you are aiming at, their published shortlisting figures from earlier years, and then aim for a comfortable margin above them. The guide to PSU and M.Tech routes explains how each route works, and how to prepare for GATE shows how to plan for a target score. The GATE exam guide covers the rest of the exam.
Using PrepMatic mocks to gauge your marks
PrepMatic mocks give you instant results with a subject-wise breakdown, the subject that cost you the most marks, and a predicted score and readiness that update as you take more tests. Treat these as a guide to your preparation, not as a GATE score: they do not reproduce the official normalisation or the top-group average. A useful habit is to compare your mock marks with last year's published qualifying mark for your paper. Questions on PrepMatic are AI-generated, so for calibration use the real papers listed in GATE previous year papers and mock tests.
Frequently asked questions
What is the GATE score out of?
The GATE score is reported on a scale that runs up to 1000. A candidate at the General qualifying mark scores 350, and one at the average of the top 0.1% scores 900.
What is the minimum qualifying mark in GATE?
The General qualifying mark is max(25, min(40, μ + σ)) out of 100, so it always lies between 25 and 40. OBC-NCL and EWS candidates need 0.9 times that, and SC, ST and PwD candidates need two-thirds of it. The figure is decided after the exam, so there is no 2027 value yet.
How is the GATE score calculated?
Your raw marks out of 100 (or normalised marks, if your paper ran in several sessions) go into Score = 350 + 550 × (M − Mq) / (Mt − Mq). The score is therefore not the same as your marks: it depends on the paper's qualifying mark and the top-group average as well.
How long is a GATE score valid?
Three years from the date of the result. For GATE 2027, that runs from 19 March 2027.
Is there normalisation in every GATE paper?
No. Normalisation applies only to papers held in more than one session. If your paper has a single session, your raw marks are used directly.
Official sources
- GATE 2027 information brochure: the score formula, normalisation method and qualifying marks (section 8).
- GATE 2027 official website (IIT Madras): schedule, results and scorecard information.
- GATE 2026 cut-off page (IIT Guwahati): the 2026 qualifying marks and candidate counts shown above.
Check the latest notice on the official site for 2027 dates, fees and scorecard rules.
Turn the GATE syllabus into a plan you can finish
Take a free GATE assessment on PrepMatic. It shows where you stand in your branch, turns the syllabus into a plan around the hours you actually have, and builds mock tests focused on your weak subjects.
Start free assessment →Free forever plan · No credit card · 2-minute setupRelated guides
- GATE exam patternGATE 2027 pattern: 65 questions, 100 marks, 3 hours, MCQ, MSQ and NAT, negative marking, marks split by paper, 30 papers and two-paper combinations.
- GATE eligibilityWho can apply for GATE 2027: degrees and year of study, no age limit or attempt cap, foreign candidates, documents, fees and the 12 October late-fee deadline.
- GATE previous year papersWhere to download GATE previous year papers (2007 to 2026) and answer keys (2021 to 2026), how to use them topic-wise and full-length, plus mock test analysis.
- After GATE: PSU and M.TechHow to use a GATE score: PSU recruitment (POWERGRID ET-2027 and the GATE 2026 list), M.Tech via COAP and CCMT, stipend, PhD fellowship and a checklist.
- How to prepare for GATEHow to prepare for GATE 2027: a plan for the 17 weeks left, plus 6-month and 12-month plans, subject order, revision, mocks and tips for working aspirants.