JAM 2026 paper at a glance
For JAM 2026, each paper was a three-hour computer-based test in English with 60 questions carrying 100 marks. All three sections were compulsory. Section A tested single-correct multiple-choice questions (MCQs), Section B used multiple-select questions (MSQs), and Section C used numerical-answer questions (NATs). The exam was conducted on 15 February 2026. The information below explains the 2026 structure as published by IIT Bombay; it is not confirmation of a later cycle’s pattern.
The 2026 paper was shown on screen in a random question sequence. Candidates answered MCQs and MSQs using a mouse and entered NAT responses through a virtual numeric keypad. A virtual calculator was available; bringing a personal calculator or other electronic device was prohibited. The test ended automatically when the allotted time expired. These instructions matter to practice: learn the interface and leave time to enter and review numerical responses, rather than assuming paper-based conditions.
Section A: MCQs and negative marks
Section A had 30 MCQs worth 50 marks in total: ten one-mark questions and twenty two-mark questions. Each question had four choices and exactly one correct choice. A wrong response to a one-mark question deducted one-third mark; a wrong response to a two-mark question deducted two-thirds. Unattempted questions earned zero.
The penalty is asymmetric: a correct answer adds one or two marks, while a wrong answer removes a fraction of one mark. That does not automatically make guessing worthwhile. Use a confidence rule during practice. If you can eliminate options and narrow the possibilities, the expected value may differ from a blind guess; if you have no basis for a choice, protect time for questions you can solve. This is a planning heuristic based on the official marking rule, not an official test-taking directive.
Train the two mark values separately. On one-mark items, avoid spending disproportionate time on a low-confidence calculation. On two-mark items, show enough working on your rough pad to catch sign, unit or algebra errors before selecting an option. During a mock, record not only correct and incorrect responses but also whether each wrong answer came from a concept gap, an avoidable calculation mistake or an unsupported guess. The marking scheme makes that distinction useful because wrong MCQs reduce the score.
Section B: MSQs reward complete selection
Section B contained ten MSQs worth two marks each, for 20 marks total. An item could have one or more correct options. The official instructions award full credit only when all correct choices and no incorrect choices are selected. There was no negative marking and no partial credit.
Do not treat an MSQ as an MCQ with extra options. Evaluate each choice independently against the question. If you select only the options you recognize but leave another correct choice unmarked, the rule provides no partial score. If one extra option is wrong, it also prevents full credit. In practice, practise explaining why every selected option is correct and why each excluded one is false. For a timed mock, flag uncertainty and return if time allows rather than turning a hunch into a complete answer.
Section C: numerical answers
Section C contained 20 NAT questions, split evenly between ten one-mark and ten two-mark items, for 30 marks. The paper did not show answer choices; candidates entered a signed real number using the on-screen keypad. There was no negative marking for an incorrect NAT and no mark for an unanswered question.
The absence of answer options changes how you check work. Practise moving from setup to a numerical result without using elimination, and verify signs, units, rounding or requested precision against the instructions for the actual item. Use the virtual keypad in official mocks so you are familiar with the entry method. The official source confirms the signed-real-number format; it does not justify assuming a particular rounding tolerance for every question, so follow the specific paper instructions.
Time and paper choice
The three-hour limit covers all 60 questions and all three compulsory sections. One simple allocation for practice is to reserve a first pass for questions you can solve confidently, then revisit flagged work and leave a final buffer to check entered answers. Treat that as a personal experiment, not a prescribed JAM time split: the official pattern does not assign separate section timers or publish a mandatory per-question pace.
JAM 2026 allowed one or two test papers only when the selected papers fell in different sessions. The official timetable placed CY, GG and MA in the forenoon, and BT, EN, MS and PH in the afternoon. A second paper may broaden eligible programme options, but it also adds another syllabus and preparation load. Check both the current session table and the target programmes’ paper mapping before choosing; the 2026 allocation is historical and may not repeat.
A focused practice loop
Use the official mock test to learn the screen, answer selection and virtual keypad. Then solve previous papers against the master question paper and its final key. The official key page cautions that the question order shown on a candidate’s console may differ from the master paper, so compare by question content rather than blindly matching position. Track section-level accuracy and the source of errors. For Section A, include negative marks in your score calculation; for Sections B and C, apply the no-negative rule and the all-or-nothing MSQ rule.
The format creates different decision costs across sections. A risky MCQ can lose marks, while a wrong MSQ or NAT does not carry a negative penalty, but MSQ scoring still requires the full correct set. Your practice review should reflect those mechanics. Once a new JAM brochure appears, verify whether question counts, timing, marking and permitted tools changed before carrying this 2026 method forward.