GATE 2027 exam pattern: 65 questions in 3 hours, MCQ + MSQ + NAT

GATE 2027 is a 3-hour computer-based test with 65 questions worth 100 marks, split across MCQ (1 of 4, with 1/3 negative marking), MSQ (multiple select, no negative) and NAT (numerical answer type, no negative), plus an on-screen scientific calculator inside the platform. General Aptitude is compulsory (10 questions, 15 marks); the remaining 50 questions are paper-specific. Same structure across all 30 paper codes.

How are the 100 marks distributed across sections?

Note: For some papers (CY, MA, PH, EY, XL, XH), Engineering Math is replaced by subject-specific content. Always check the official syllabus for your paper code.

Question types

How does the raw mark become a GATE score out of 1000?

Your raw mark (out of 100) is normalised to a GATE score on a 0-1000 scale using this formula:

GATE Score = 350 + 250 × (M − Mq) / (Mt − Mq)

where M is your raw mark, Mq is the qualifying cutoff for your paper (typically ~25 for General), and Mt is the average of the top 0.1% of candidates in your paper. The qualifying cutoff is the floor (GATE score ~350); the top 0.1% mean is the ceiling (GATE score ~1000).

The 30 GATE 2027 paper codes

GATE 2027 will conduct 30 papers across engineering, science, architecture, humanities and economics. Each paper has its own subject-specific section but shares the same 10-question General Aptitude section. The high-volume engineering papers (CS, ME, EE, EC, CE) carry the bulk of the candidate pool; specialist papers like Mining, Petroleum or Naval Architecture see a few thousand candidates each.

The 1-mark vs 2-mark question mix

Inside the 65-question paper, marks per question are not uniform. Each paper has a mix of 1-mark and 2-mark questions, and the conducting body publishes the breakup for each section in the information brochure. The standard split sums to 100 marks:

Two-mark questions are typically more complex - longer derivations, multi-step numerical computations, or compound MCQs that test two concepts inside one stem. One-mark questions are usually single-concept and faster. The implication for attempt strategy: a 2-mark question is worth twice as much as a 1-mark question but typically takes ~1.5x the time, so the marks-per-minute return is actually higher on 2-mark questions you can solve confidently.

Negative marking, the actual math

Negative marking on GATE only applies to MCQ-type questions. There is no negative marking on MSQ or NAT. The penalty is proportional to the marks of the question.

• 1-mark MCQ wrong: -1/3 mark (~0.33 negative).
• 2-mark MCQ wrong: -2/3 mark (~0.67 negative).
• MSQ wrong / partially wrong: 0 marks. No negative. But you only get the full marks if every correct option is selected and no incorrect option is selected; partial selection earns 0.
• NAT wrong: 0 marks. No negative.

Work the expected-value math on a random MCQ guess. With 4 options, random pick has a 25% chance of being correct and 75% of being wrong. On a 1-mark MCQ that is 0.25 x (+1) + 0.75 x (-1/3) = 0. So a fully random guess is exactly break-even on expectation. But the moment you can rule out 1 option, the math becomes 0.33 x (+1) + 0.67 x (-1/3) = +0.11; with 2 options ruled out, +0.33. This is why the standard rule is: only attempt an MCQ when you have eliminated at least 1 option on substantive grounds.

MSQ and NAT: why they are scored differently

MSQ (Multiple Select Question) and NAT (Numerical Answer Type) were introduced to reduce the role of guessing on GATE and reward candidates who can compute / reason from first principles rather than test-take their way through MCQ elimination. Two things follow from that intent.

• MSQ has no partial credit: The options labelled correct on the answer key form a set. You only score the full marks if the set you select is identical - every correct option in, every incorrect option out. Get one wrong and you score 0. There is no negative, but no consolation either.
• NAT is a typed numeric answer: You enter a number into a virtual keypad. The conducting body accepts a tolerance range (e.g. 23.5 +/- 0.1) for floating-point answers. Match the value within tolerance and you score the full marks; outside tolerance and you score 0.

The practical takeaway: MSQ and NAT are not riskier than MCQ; they are differently structured. Random selection on MSQ rarely matches the exact correct set, so do not attempt blind. NAT cannot be guessed at all - either you have the answer or you do not. But because there is no negative, attempting every NAT you can compute is a free shot at the marks.

The on-screen scientific calculator

GATE 2027 provides a virtual scientific calculator inside the test platform. No physical calculator (programmable, non-programmable, scientific or basic) is allowed in the hall. The virtual calculator is a non-programmable model that supports:

• Arithmetic: addition, subtraction, multiplication, division, modulo.
• Powers and roots: x squared, x cubed, x to the y, square root, n-th root.
• Trigonometry: sin, cos, tan and their inverses, in radians and degrees.
• Logarithm and exponential: log base 10, natural log, e to the x.
• Memory: M+, M-, MR (recall), MC (clear), MS (store).
• Constants: pi, e.
• Hyperbolic and combinatorial functions on extended panels.

The calculator opens in a floating window that overlaps the question text. The two UX gotchas: keyboard input does not type values into the calculator (every character is a click on the rendered keypad), and the display does not show the full expression - only the current value, which makes long chained computations error-prone. Practise on the actual virtual calculator before exam day; do not rehearse on a physical device and assume the interface will translate.

How should you budget the 3 hours across 65 questions?

You have 3 hours for 65 questions - average of ~2.75 minutes per question. But NAT questions take longer (~4-5 min) and aptitude takes less (~1 min). A typical split:

• First 30 minutes: All 10 GA questions + first pass on Engineering Math. Bank easy marks.
• 90 minutes: Core subject MCQs and MSQs. Skip NATs on first pass.
• Final 50 minutes: NAT questions + marked-for-review revisit. NATs reward careful calculation; rushing costs marks.
• Last 10 minutes: Final review of flagged questions. Resist the urge to guess on MCQs you skipped - negative marking is real.

Per-paper time allocation cheat sheet

The 65-question, 3-hour budget has the same shape for every paper, but the internal distribution of NAT-heavy vs concept-heavy questions varies. Treat the allocations below as a starting baseline; adjust based on your own mock data.

An MCQ + MSQ + NAT attempt rulebook

• Always attempt every NAT you can solve fully. No negative; the only downside is time spent.
• Attempt MSQ only when you are sure of the full correct set. Partial confidence on MSQ leads to a 0 because no partial credit is awarded.
• Attempt an MCQ when you have ruled out at least 1 option. Random 4-way guess is break-even; 3-way is positive expected value.
• Skip outright on completely cold MCQs. An unattempted MCQ scores 0; a random wrong answer costs you 1/3. The score-maximising move on a cold MCQ is to skip.
• Flag-for-review aggressively early.Use the platform's mark-for-review button on anything you are not 100% confident on; come back in the final 30 minutes with a clean head.