IITM BS Qualifier Exam 2027: Syllabus, Cut-offs & Prep Plan

The IITM BS qualifier exam is a four-hour, in-person exam on the first four weeks of four courses: Mathematics for Data Science I, Statistics for Data Science I, Computational Thinking and English I. Before you can sit it, you study those four weeks online and clear a minimum score on the weekly assignments. To qualify, a General-category candidate needs at least 40% in every course and at least 50% on average. Clear it, and you can register for the Foundation level of the IIT Madras BS.

That's the whole process in one paragraph. This guide is the long version: how each gate works, what's actually in the syllabus (with solved examples from past papers), a week-by-week plan, how to use previous-year papers properly, and the exam-day decisions that separate a comfortable pass from a near miss.

All cut-offs and fees below are from the official admissions pages for the Management and Data Science and Data Science programmes, checked in October 2026 for the January 2027 batch. If you're new to the programme itself, start with our complete guide to the IIT Madras BS degree.

The qualifier at a glance#

How the qualifier process works#

The qualifier isn't a one-day entrance test. It's a short, real slice of the programme, and the best way to think about it is as your first month as an IITM BS student, with an exam at the end.

1. You apply and pay the application fee.
2. Week 1 opens. You get lecture videos, notes and practice activities for all four courses, plus one graded assignment per course.
3. Weeks 2, 3 and 4 follow the same pattern, one week at a time.
4. Gate 1 is checked: the average of your best two of the first three assignment scores, in every course.
5. If you clear Gate 1, you're eligible for the qualifier exam, held in person at your chosen exam city.
6. Gate 2 is the exam result: a minimum in each course and a minimum average.
7. If you clear both, you register for Foundation courses, and your qualifier score carries forward as Quiz 1 for the Foundation courses you take in the very next term.

Gate 1: the assignment cut-off#

For each of the four courses, IIT Madras takes your best two of the first three weekly assignment scores and averages them. That average must reach the minimum for your category:

Three practical consequences:

• Week 4's assignment doesn't count towards Gate 1. It's still worth doing, because it's the best practice you'll get for the exam, but your eligibility is decided by the first three.
• One bad week is survivable. Because only your best two of three count, a missed or poor assignment in one week can be covered by the other two.
• Two bad weeks usually aren't. If you score poorly in two of the first three assignments for any course, you're out of this attempt, and you'll have to apply again as a new applicant and pay the full fee.

Assignments are open-book and you have days to complete each one. There's no good reason to let this gate be the one that stops you. Aim for 80% or more, not 40%, so a bad week doesn't matter.

Gate 2: the exam cut-off#

Both conditions must hold. This is the single most misunderstood rule of the qualifier, so let's make it concrete.

Candidate A is excellent at maths and statistics and scores 85, 80 and 75 in Maths, Stats and English. But they skimmed CT, assuming it would be easy, and scored 38. Their average is 69.5%, far above 50%. They don't qualify.

Candidate B is nobody's idea of a topper: 52, 55, 60 and 48. Their average is 53.75%. Every course is above 40%. They qualify.

The lesson writes itself. Your preparation should be organised around your weakest course. A balanced 55% everywhere beats a brilliant 80% average with one course in the danger zone.

What it costs#

An extra exam facilitation fee applies if you take the qualifier exam outside India. Application fees aren't refundable, even if you miss the exam.

Who can skip it#

Candidates who were eligible to appear for JEE Advanced in the relevant year (the Management and Data Science page lists 2025 or 2026) can take the direct-entry route: pay the admission fee, submit JEE proof, and, once it's verified, register for Foundation courses directly. If that's you, you don't need this guide, though the syllabus section below is still a useful preview of your first term.

The exam format#

The qualifier exam is in person and invigilated, at an exam centre in the city you chose. It lasts four hours and covers all four courses. You'll see the same question formats that IITM uses throughout the programme:

You can move between courses during the exam. That freedom is a trap if you don't plan for it: four hours feels like plenty until you've spent 80 minutes fighting one maths question. We cover time strategy later in this guide.

The syllabus, course by course#

The qualifier covers weeks 1–4 of each course. Below is what each course asks for, the patterns that repeat across past papers, and solved examples. Every example marked "PYQ" is modelled on a real previous-year qualifier question, and the answers were checked against the official keys used in BTechi's cheat notes.

Mathematics for Data Science I#

Weeks 1–4: sets, relations and functions; coordinate geometry and straight lines; quadratic functions; polynomials.

The course rewards precise definitions. Many questions are less about calculation than about whether you know exactly what a term means: is this relation transitive? Is this function one-to-one on this domain?

Patterns that repeat:

• Venn diagram counting with two or three sets, often phrased as a word problem about customers, students or sports.
• Properties of relations: reflexive, symmetric, transitive, and counting the elements of a relation defined by a rule.
• Functions: one-one, onto, domain and range, especially when the domain is restricted.
• Straight lines: slope, intercepts, parallel and perpendicular lines, angle between lines, distance from a point to a line, intersection points.
• Quadratics: vertex, maximum or minimum, roots, discriminant, and word problems about profit or area.
• Polynomials: remainder and factor theorem, and reading graphs from roots and multiplicities.

Solved example 1 (PYQ pattern, Sep 2023). In a group of 60 people, 40 like cricket and 35 like football. (a) What is the minimum number who like both? (b) If 10 like neither, how many like only cricket?

• (a) The overlap is smallest when as few people as possible like neither. With nobody outside both sets, both = 40 + 35 − 60 = 15.
• (b) If 10 like neither, 50 like at least one. Both = 40 + 35 − 50 = 25, so only cricket = 40 − 25 = 15.

Solved example 2 (PYQ pattern, Jan 2024). Of 100 customers, 50 bought milk, 40 bread and 30 fruits. 20 bought milk and bread, 15 milk and fruits, 10 bread and fruits, and 5 bought all three. How many bought milk and fruits but not bread? How many bought none?

• Milk and fruits but not bread = 15 − 5 = 10.
• At least one = 50 + 40 + 30 − 20 − 15 − 10 + 5 = 80, so none = 20.

The habit that makes these easy: fill the Venn diagram from the centre outwards, starting with "all three".

Solved example 3 (PYQ, May 2024). For y = x² + 4x + 12, find the vertex, the minimum value, the x-intercepts and the range.

• Vertex at x = −b/2a = −4/2 = −2; y = 4 − 8 + 12 = 8. Vertex (−2, 8), minimum value 8.
• Discriminant = 16 − 48 < 0, so no x-intercepts.
• The parabola opens upwards from 8, so the range is [8, ∞).

Solved example 4 (PYQ pattern, Jan 2024). Ram copied the constant term of a quadratic wrongly and got roots 3 and 4. Shyam copied the x-coefficient wrongly and got roots 1 and 10. What are the correct roots?

• A wrong constant term doesn't change the sum of roots, so the true sum is 3 + 4 = 7.
• A wrong x-coefficient doesn't change the product, so the true product is 1 × 10 = 10.
• The equation is x² − 7x + 10 = 0, with roots 2 and 5.

This one is a favourite because it tests whether you understand why the sum and product formulas work, not just the formulas.

Solved example 5 (PYQ pattern, May 2023). A firm's profit is P(x) = −2x² + 24x − 40, where x is tonnes of raw material. For what x is profit maximised, and when is profit zero?

• Maximum at x = −24 / (2 × −2) = 6; P(6) = −72 + 144 − 40 = 32.
• Zero profit when −2x² + 24x − 40 = 0, i.e. x² − 12x + 20 = 0, so x = 2 or x = 10.

More solved Maths I examples#

PYQ (May 2023). The Cartesian product A × A has 9 elements. Two of them are (2, 0) and (0, 8). Find the sum of all the elements of A.

• |A × A| = |A|², so |A|² = 9 and |A| = 3.
• The pairs (2, 0) and (0, 8) tell us 2, 0 and 8 are all in A, so A = {0, 2, 8}.
• Sum = 10.

Relation counting. S = {1, 2, 3, 4} and R = {(a, b) : a divides b}. How many pairs are in R, and how many pairs of S × S are not in R?

• a = 1 divides everything: 4 pairs. a = 2: (2, 2) and (2, 4). a = 3: (3, 3). a = 4: (4, 4).
• |R| = 8. |S × S| = 16, so 8 pairs are not in R.

Writing the relation out as a list of pairs, then testing properties against the list, is far more reliable than reasoning from the rule.

PYQ (May 2024). Line L₁ passes through (3, −2) and has x-intercept 1. Line L₂ passes through (−1, 5) and has y-intercept −1. Find tan θ, where θ is the angle between them.

• L₁ passes through (1, 0) and (3, −2): slope m₁ = −2/2 = −1.
• L₂ passes through (0, −1) and (−1, 5): slope m₂ = 6/(−1) = −6.
• tan θ = |(−1 − (−6)) / (1 + 6)| = 5/7.

The key step is turning "x-intercept 1" into the point (1, 0). Intercept information is always a point in disguise.

Best-fit line. For the points (1, 2), (2, 3) and (3, 5), which line fits better: y = x + 1 or y = 1.5x + 0.5?

• y = x + 1 predicts 2, 3, 4. Errors 0, 0, 1. Sum of squared errors = 1.
• y = 1.5x + 0.5 predicts 2, 3.5, 5. Errors 0, −0.5, 0. Sum of squared errors = 0.25.
• The second line fits better (SSE 0.25 < 1).

Remainder and factor theorem. Find the remainder when p(x) = x³ − 2x² + 4x − 5 is divided by (x − 2).

• p(2) = 8 − 8 + 8 − 5 = 3.

And if (x − 1) is a factor of x³ + kx² − 4x + 2, then p(1) = 1 + k − 4 + 2 = 0, so k = 1.

Reading a polynomial graph. Describe the graph of p(x) = (x + 2)(x − 1)².

• Degree 3 with a positive leading coefficient: it starts low on the left and ends high on the right.
• At x = −2 (power 1) the graph crosses the x-axis; at x = 1 (power 2) it touches and bounces.
• The y-intercept is p(0) = 2 × 1 = 2.

Quick formulas to have cold:

The traps: forgetting that a relation containing (a, b) and (b, a) must also contain (a, a) and (b, b) to be transitive; treating x² as one-to-one on all real numbers; confusing the slope of a line with its intercept form; and, in polynomial graphs, forgetting that a root of even multiplicity touches the x-axis without crossing it.

The full set of formulas, traps and worked examples for weeks 1–4 is in our Maths I cheat notes.

Statistics for Data Science I#

Weeks 1–4: types of data and scales of measurement; describing categorical data; describing numerical data; association between two variables.

Stats I in the qualifier is less about heavy calculation and more about choosing the right summary and following the course's conventions. It's very gettable, provided you learn those conventions.

Patterns that repeat:

• Classifying variables: categorical or numerical, discrete or continuous, and the scale (nominal, ordinal, interval, ratio).
• Categorical summaries: frequency tables, relative frequencies, bar and pie charts, mode and (for ordinal data) median.
• Numerical summaries: mean, median, mode, range, variance, standard deviation, percentiles and quartiles, plus the effect of adding or correcting values.
• Association: contingency tables, scatter plots, covariance and correlation.

Solved example 1 (PYQ pattern, Sep 2023). A pie chart of 500 fruits sold shows mangoes as a 72° slice. How many mangoes were sold?

• 72/360 = 1/5 of the total, so mangoes = 500 × 1/5 = 100.

Solved example 2 (PYQ pattern, Sep 2023). Sunidhi buys T-shirts of sizes M, L, S, XL, S, and then one more M. What happens to the mode and the median?

• Before: S, S, M, L, XL. Mode S; median (3rd value) M.
• After: S, S, M, M, L, XL. S and M both appear twice, so the data is bimodal (S and M). The median is the 3rd and 4th values, M and M, so it stays M.

Notice that this only works because T-shirt sizes are ordinal. For nominal data, like city names, a median makes no sense at all.

Solved example 3 (our own, in the style of the course). The mean of 20 observations is 45. Later, it's found that one value recorded as 27 should have been 72. What's the correct mean?

• Original total = 20 × 45 = 900.
• Corrected total = 900 − 27 + 72 = 945.
• Correct mean = 945 / 20 = 47.25.

"Wrong value corrected" questions appear in some form in almost every paper. Always go through the total, never try to adjust the mean directly.

More solved Stats I examples#

PYQ pattern (Sep 2023). The average annual package in a batch is ₹8 lakh. Every package has 5 subtracted from it and is then doubled. What's the new average?

• If y = 2(x − 5), then the mean of y = 2 × (mean of x − 5) = 2 × 3 = ₹6 lakh.
• The order of operations matters: 2(x̄ − 5) is not 2x̄ − 5.

Transformations and spread. Marks have mean 40 and sample variance 25. Each mark has 5 subtracted and is then doubled. Find the new mean, variance and standard deviation.

• Mean: 2(40 − 5) = 70.
• Variance: subtracting a constant changes nothing; doubling multiplies variance by 2² = 4. New variance 100.
• SD = √100 = 10.

PYQ (May 2023): the corrected standard deviation. Ten observations have mean 22 and sample variance 81. One value, 20, was recorded as 15. Find the correct sample standard deviation.

• Σx = 10 × 22 = 220. Corrected: 220 − 15 + 20 = 225, so the new mean is 22.5.
• Σx² = (n − 1)s² + n x̄² = 9 × 81 + 10 × 22² = 729 + 4,840 = 5,569. Corrected: 5,569 − 15² + 20² = 5,744.
• New s² = (5,744 − 10 × 22.5²) / 9 = 681.5 / 9 ≈ 75.72.
• s ≈ 8.70 (the official key accepted 8.68–8.71).

This is one of the hardest calculations in the qualifier, and it rewards one habit above all: work with totals (Σx and Σx²), not with the mean and variance directly.

PYQ (Sep 2023): IQR and outliers. Find the IQR of 14, 20, 40, 23, 18. Are there outliers?

• Sorted: 14, 18, 20, 23, 40 (n = 5).
• Using the course's method: Q₁ position = 0.25 × 5 = 1.25, round up to 2, so Q₁ = 18. Q₃ position = 0.75 × 5 = 3.75, round up to 4, so Q₃ = 23.
• IQR = 5. Fences: 18 − 7.5 = 10.5 and 23 + 7.5 = 30.5, so 40 is an outlier.

Note the rounding rule: the position is rounded up even when the decimal is small (1.25 becomes 2). Calculators and spreadsheets often use other methods and give different quartiles, which is why it's worth practising the IITM method by hand.

PYQ pattern (May 2024): contingency tables. Of 200 people, 120 are male and 80 female. 150 own a smartphone, and 30 of the females don't. How many males own a smartphone?

• Females who own one = 80 − 30 = 50.
• Males who own one = 150 − 50 = 100.

Draw the empty two-by-two table with totals first, write in every number you're given, then fill the blanks so rows and columns add up.

Covariance. For x = 1, 2, 3 and y = 2, 4, 9, find the population and sample covariance.

• Means: x̄ = 2, ȳ = 5. Products of deviations: (−1)(−3) + (0)(−1) + (1)(4) = 7.
• Population covariance = 7/3 ≈ 2.33. Sample covariance = 7/2 = 3.5.

Questions ask for both kinds, so read the word "population" or "sample" carefully.

PYQ pattern (Jan 2023): perfect correlation. Deepak scores exactly twice Mayur's marks in every subject. What can you say?

• Y = 2X is a rising straight line, so the correlation r = 1, cov(X, Y) = 2 Var(X), Var(Y) = 4 Var(X), and SD(Y) = 2 SD(X).
• Remember that r measures how tightly points hug a line, not how steep the line is.

The classification questions people get wrong:

Temperature in Celsius is the classic: 0 °C isn't "no temperature", and 20 °C isn't twice as hot as 10 °C, so it's interval, not ratio.

The traps: using the mean for categorical data coded as numbers; using a median for nominal data; forgetting that a dataset can have more than one mode; and, in percentile questions, using a method from school instead of the course's method. Our Stats I cheat notes spell out the IITM percentile method that the official answer keys follow.

Computational Thinking#

Weeks 1–4: variables, iteration and filtering; datasets represented as cards and tables; flowcharts and pseudocode; procedures and multiple filters.

CT is the course that most often decides who qualifies. It looks friendly, because there's no real programming language, but it tests exact tracing: following a procedure step by step and keeping track of every variable. There's no partial credit for "nearly".

The datasets. CT questions are built on a small set of datasets represented as cards: student scores (with gender, city and marks in several subjects), words from a paragraph (with part of speech and letter count), shopping bills (with shop, customer and items), and a few others. Learn their fields early; questions assume you know them.

Patterns that repeat:

• Count and sum in one pass: iterate over the cards, filter on a condition, keep a count or a running total.
• Max and min, and "who has it": the highest or lowest value, often with a filter, and what happens with ties.
• AND and OR filters: how many cards satisfy a combined condition.
• Procedures: what a procedure returns for a given input, and what a mystery procedure computes.
• Slab and bill problems: if/else-if chains for tariffs.
• Nested iterations: comparing every pair of cards, and how many comparisons that takes.
• "What does A represent?": the hardest category: work out what a variable holds at the end of a pseudocode block.

Solved example 1 (PYQ pattern). In the Scores dataset, 14 students are female, 12 are from Chennai, and 5 are both female and from Chennai. How many cards pass the filter X.Gender == "F" or X.CityTown == "Chennai"?

• OR counts everyone in either group, minus the overlap counted twice: 14 + 12 − 5 = 21.

Solved example 2 (PYQ pattern, May 2023). Electricity costs ₹5 per unit for the first 100 units, ₹7 per unit for the next 100, and ₹10 per unit after that. What's the bill for 250 units?

• First 100 units: 100 × 5 = ₹500.
• Next 100: 100 × 7 = ₹700.
• Remaining 50: 50 × 10 = ₹500.
• Total = ₹1,700.

The trap is charging all 250 units at ₹10. Slab pricing applies each rate only to the units inside its slab.

Solved example 3 (pairs). The Scores dataset has 30 cards: 10 from Chennai, 8 from Vellore and 12 from Madurai. How many pairs of students are from the same city? From different cities?

• Same city: C(10, 2) + C(8, 2) + C(12, 2) = 45 + 28 + 66 = 139.
• All pairs: C(30, 2) = 435, so different cities = 435 − 139 = 296.

Solved example 4 (binning). You compare every pair of 30 cards. How many comparisons is that, and how many if you first split the cards evenly into 3 bins and only compare within a bin?

• Without binning: C(30, 2) = 435.
• With binning: 3 × C(10, 2) = 3 × 45 = 135.

Binning appears in CT precisely because it's the intuition behind efficient algorithms: you avoid comparisons you know can't matter.

Solved example 5 (tracing). Run Euclid's GCD loop (while b ≠ 0: (a, b) = (b, a mod b)) with a = 36 and b = 24.

The loop stops when b = 0, so the answer is 12. Writing the table is what makes tracing reliable. Do it every time, even when you think you can see the answer.

More CT patterns worth knowing#

Two passes when you need an average first (PYQ, Jan 2024). "How many students scored above the class average?" can't be answered in one pass, because you don't know the average until you've seen every card. The pseudocode reads the cards once to compute the average, puts them back, and then reads them again to count:

sum = 0
n = 0
while (Table 1 has more rows) {
    Read the first row X in Table 1
    sum = sum + X.Total
    n = n + 1
    Move X to Table 2
}
avg = sum / n
Move all rows from Table 2 to Table 1
count = 0
while (Table 1 has more rows) {
    Read the first row X in Table 1
    if (X.Total > avg) {
        count = count + 1
    }
    Move X to Table 2
}

Two details examiners like to test: forgetting to move the rows back to Table 1 (the second loop would then run zero times), and > versus ≥ (is a student exactly at the average counted?).

Digit loops. "What is the sum of the digits of 4096 using the mod-10 loop?" Repeatedly take n mod 10 (the last digit), add it to the total, and replace n with the integer division n / 10: 6 + 9 + 0 + 4 = 19. The same loop, with small changes, counts digits, reverses a number, or checks for a palindrome.

Keeping the top three in one pass. A common question asks for the first, second and third highest totals. The pattern: when a new total beats first place, shift first to second and second to third before updating. Getting the order of those assignments wrong overwrites a value you still need, and tracing with a table is the only reliable way to catch it.

The traps: off-by-one errors in loop bounds; forgetting to reset a variable inside an outer loop; ties in max/min questions (does the first card or the last card win? It depends on whether the comparison is > or ≥); and confusing "number of cards" with "number of pairs". The CT cheat notes list every repeating pattern with its trick.

English I#

Weeks 1–4: sounds and pronunciation; parts of speech and grammar; vocabulary and usage; reading comprehension.

English is the course people either ignore ("I speak English fine") or panic about ("phonetics?!"). Neither is warranted. It's very learnable in four weeks if you study it a little every day.

Patterns that repeat:

• Vowel sounds: which word has a particular vowel sound, which word has a diphthong, odd one out by sound.
• Consonants: odd one out by initial sound, often decided by where in the mouth the sound is made.
• Parts of speech: identify the noun, pronoun, verb, adjective or adverb in a sentence; abstract versus collective nouns.
• Articles and prepositions: fill in the blank with a, an, the, or the right preposition.
• Tenses and agreement: choose the correct verb form.
• Idioms, phrasal verbs and collocations: what an idiom means; which words naturally go together; matching exercises.
• Reading comprehension: questions on a short passage.

Solved example 1 (PYQ answers, 2023–24). Which word in each group has no diphthong?

• Snout / Stout / Sport / Shout → Sport (the others all have the "ow" diphthong).
• Mouth / South / Pout / Moot → Moot (a long "oo", not a diphthong).

Solved example 2 (PYQ answers, articles). Fill in the blanks: "There is __ emergency exit." "This is __ best chocolate cake." "__ Sun rises in the east." "Tomorrow will be __ cold day."

• an (before a vowel sound), the (superlatives take "the"), The (unique objects), a (one of many possible days).

The rule people forget: a and an depend on the sound, not the letter. It's "an hour" and "a university".

Solved example 3 (PYQ answers, odd one out by initial sound). near, dear, tear, gear → gear: the others start with sounds made with the tongue behind the teeth, while "g" is made at the back of the mouth.

More English I patterns#

Phrasal verbs: separable or inseparable? Put "it" in the middle. "Sum it up" works, so sum up is separable. "Come it across" doesn't, so come across is inseparable. Past papers have asked about sum up, call by and come across, among others.

Idioms that have appeared in past papers include a dark horse (an unexpected winner), cut corners (do something the cheapest or quickest way), break the ice (ease tension at a first meeting), get out of hand (become hard to control), cost an arm and a leg (be very expensive) and keep someone in the loop (keep them updated). Learn idioms in small groups of five a day, and use each in a sentence of your own.

Collocations. Almost every paper has a five-item matching question on words that naturally go together: weather forecast, bank account, unanimous decision, strong coffee, short temper, job offer. Match in two passes: lock the pairs you're certain of first, and the remaining options usually fall into place.

Homophones and confusables. week/weak, principal/principle, stationary/stationery, affect/effect. These are cheap marks if you've seen them and expensive if you haven't.

The traps: spelling-based answers to sound-based questions (English spelling is a poor guide to pronunciation); treating "Parliament" as an abstract noun (it's collective); and reading comprehension answers based on what you think rather than what the passage says. The English I cheat notes collect the sound charts and the most-asked items.

Test yourself: a 12-question warm-up#

These are original questions in the style of the qualifier, three per course. Try them on paper, without a calculator, before reading the answers. If you get fewer than two right in any course, that's your starting point.

Maths I

1. A = {1, 2, 3} and B = {a, b}. How many functions are there from A to B? How many of them are onto?
2. Find the equation of the line through (2, 3) that is perpendicular to 2x − y + 4 = 0.
3. For which values of k does x² + kx + 16 = 0 have real roots?

Stats I

4. For the data 4, 7, 7, 9, 13, find the mean, median, mode and range.
5. A survey records customer satisfaction as 1 to 5 stars. What type of variable is it, and on what scale?
6. Find the population standard deviation of 2, 4, 6.

Computational Thinking

7. A procedure counts cards in Scores where X.Maths > 80 and X.Physics > 80. If 18 students have Maths above 80, 15 have Physics above 80 and 9 have both, what does it return?
8. Using the slabs ₹5 per unit for the first 100 units, ₹7 for the next 100 and ₹10 after that, what's the bill for 180 units?
9. How many comparisons does it take to compare every pair among 12 cards, each pair exactly once?

English I

10. Fill in the article: "She is __ honest person."
11. Which word contains a diphthong: boat, bought or boot?
12. What does once in a blue moon mean?

A week-by-week preparation plan#

This plan assumes about 2–3 hours a day, five or six days a week, which is what most people with a college timetable or a job can manage. If you have more time, don't add hours; add more practice questions.

Before week 1: the warm-up (optional, one week)#

• Revise Class 10 algebra: linear equations, factorising, the quadratic formula.
• Read our four cheat notes once each, slowly, without trying to memorise.
• Try one previous-year paper untimed, simply to see what the questions look like. Don't score it; note what felt unfamiliar.

Week 1: build the system#

Week 2: same rhythm, plus past papers#

Repeat the week 1 rhythm with week 2 content. On day 6, add the week 1–2 questions from one previous-year paper for each course. Do them untimed, check them against the key, and add mistakes to your log.

Week 3: the gate week#

This week's assignment completes Gate 1. Before submitting, check your scores from weeks 1 and 2: if either is low in any course, week 3 is your chance to fix it. Give that course extra time this week, even at the expense of a course you're already comfortable in.

On day 6, do the week 1–3 questions from a second past paper, now timed.

Week 4: from learning to performing#

Preparing from your background#

The same four weeks feel very different depending on where you're starting from. Here's how to adjust the plan for the backgrounds we see most often.

If you're from commerce or arts#

You may not have done maths since Class 10, and that's the biggest adjustment. The good news: Maths I starts with sets and functions, which reward careful reading more than mathematical speed, and Stats I is closer to commerce mathematics than you might expect.

• Do the optional warm-up week, and spend most of it on Class 10 algebra: linear equations, factorising and the quadratic formula.
• In Maths I, write out every worked example yourself before watching the solution. Watching feels like learning; writing actually is.
• In Stats I, your comfort with tables and percentages is an advantage. Lean on it.
• Give CT daily attention from day 1. Logical tracing is new for most people, regardless of stream.

If you're an engineering student or graduate#

Maths I will feel familiar, which is its own risk: it's easy to skim lectures and then lose marks on the course's specific definitions and conventions.

• Don't skip lectures because the topic looks easy. Watch at 1.5× speed instead, and note every definition the instructor states precisely.
• Spend your saved time on English I (phonetics) and on the IITM percentile method in Stats I, which differs from what most engineering courses teach.
• Do CT tracing on paper, not in your head. Programmers are especially prone to "seeing" what code does and missing the off-by-one detail.

If you're still in school (Class 11 or 12)#

You're balancing the qualifier with board preparation, which is hard, but you have advantages: maths is fresh, and your pass stays valid for six terms.

• Pick a qualifier term that avoids your board-exam months if you can.
• Use the qualifier to reinforce school maths, not compete with it: sets, functions, straight lines and quadratics are all board topics.
• Keep the daily time small but consistent: an hour a day for four weeks is enough if you never skip.

If you're working full-time#

Your constraint is time, not ability. The plan below is built for 1–1.5 hours on weekdays.

In week 4, take one or two days of leave if you possibly can, and use them for full timed mocks. It's the single most valuable use of leave in the whole process.

If you're returning to study after a long break#

Expect the first week to feel slow and uncomfortable. That's normal and it passes. The programme was explicitly designed to be open to learners irrespective of age or academic background, and returning learners often outperform recent school leavers in the long run because they study more deliberately.

• Start the warm-up a week early.
• Learn the platform basics on day 1: where lectures, assignments and forums live, and how deadlines are displayed.
• Join the live sessions. Asking a question you think is basic is exactly what they're for.

Resources: what to use, and what to skip#

There's no shortage of qualifier material online, and most of the problem is too much of it. Here's a short, opinionated list.

Use, in this order:

1. The official course content for your term: lectures, notes, practice and graded assignments. Questions follow its conventions.
2. Previous-year qualifier papers with official keys. BTechi's PYQ library lets you attempt them in an exam-style portal.
3. Condensed revision notes. Our cheat notes for all four courses have formula-only and traps-only views for the last week.
4. Full timed mocks in week 4, such as BTechi's mock qualifiers.
5. Live sessions and the course discussion forum, for anything you're stuck on for more than 20 minutes.

Skip or limit:

• Generic textbooks for the qualifier weeks. They're excellent later; for four weeks of specific content, they add volume, not marks.
• Random YouTube playlists that "cover the whole qualifier in 6 hours". Watching isn't practising.
• Large chat groups sharing answer screenshots. Beyond the academic-integrity risk, copying assignment answers leaves you unprepared for an exam where nobody can help you. The official FAQ also notes there are no official WhatsApp or Telegram groups.

AI tools can be useful for explaining a concept in different words or generating extra practice questions in the style of the course. Use them to understand, never to complete graded assignments. You'll be alone with the questions in the exam hall.

Looking after yourself during qualifier month#

Four weeks of four courses on top of everything else in your life is a real load. A few things make it sustainable.

• Protect your sleep, especially in week 4. Memory consolidates during sleep. The marks you lose to tired, careless mistakes are almost always larger than the marks you gain from an extra late night.
• Keep the daily habit, even on bad days. On a day when you can't study for two hours, do twenty minutes. Missing a day is fine; missing three in a row is how people fall behind.
• Measure progress by mistakes fixed, not hours spent. A short session that clears three entries from your mistakes log is a good session.
• Expect one bad week. Almost everyone has one. Because Gate 1 takes the best two of three assignments, the system is designed to absorb it.
• Prepare the logistics early. Knowing where your exam centre is, how you'll get there and what you need to carry removes a surprising amount of exam-week anxiety.

How to use previous-year papers properly#

Past papers are the best preparation resource there is, and most people waste them. Here's how to get full value. (We've collected the qualifier papers with official keys in BTechi's PYQ library, which you can attempt in an exam-style portal.)

Don't start with full papers. In weeks 1–3, use past papers topic by topic: after finishing week 2 content, do only the week 1–2 questions. This tells you whether you've learned the material while there's still time to fix gaps.

Do at least two full papers under exam conditions in week 4. Four hours, in one sitting, with a clock visible. The point isn't the score; it's to find out how your attention and timing hold up in hour three.

Score by course, not overall. An overall 62% tells you nothing useful. "Maths 70, Stats 75, English 68, CT 41" tells you exactly where your next three days should go.

Review every question, including the ones you got right. For each correct answer, ask: was I sure, or did I guess? Guesses that happened to be right are future mistakes.

Spot the repeats. After two or three papers, you'll notice question types that appear every time: Venn diagrams, the T-shirt median, slab bills, diphthongs, collocation matching. Those are your guaranteed marks; make sure you'll never drop them.

Exam-day strategy#

Plan your four hours before you walk in#

Decide in advance roughly how long each course gets and in what order you'll attempt them. A simple, effective plan:

Adjust to your own profile, but don't leave your weakest course for last. That's the course where the per-course minimum lives, and you want your best hour on it, not your most tired one.

The two-pass method#

On each course, go through every question once and answer only the ones you can do confidently in under three minutes. Flag the rest. Then do a second pass on the flagged questions. This guarantees you never lose easy marks at the end of a section because you were stuck on question 4.

Multiple-select questions#

Treat each option as a separate true/false statement. Don't look for "the pattern"; examiners write MSQs specifically to punish pattern-guessing. If an option involves a calculation, do the calculation.

Numerical-answer questions#

Before you type, check: the unit asked for, the rounding instruction (decimal places), and the sign. A correct method with the answer rounded wrongly is still wrong.

When you're stuck#

Move on after three to four minutes. Come back in your second pass with fresh eyes. In CT, if a tracing question is taking forever, write the variable table: it's slower for the first two steps and much faster after that.

Logistics#

Check your exam centre location before the day and plan to arrive well within the reporting window. Carry exactly what your hall ticket asks for, usually the hall ticket itself and a valid photo ID. Read the on-screen instructions before starting the timer in your head.

Mistakes that cost people the qualifier#

These are the patterns we see most often among people who didn't clear it the first time:

1. Skipping CT because it "doesn't have real coding", then scoring under 40.
2. Starting late. Treating the first week as optional, then trying to catch up on three weeks of four courses at once.
3. Aiming for 40% in assignments. It leaves no margin, and the habit carries into the exam.
4. Learning formulas without conditions. Knowing −b/2a but not when the vertex is a maximum.
5. Studying from random sources. The course has its own conventions (percentiles, pseudocode syntax). Study the official material first, then use summaries like our cheat notes to revise.
6. Never doing a full timed paper. Four hours is long. Stamina is a skill.
7. Ignoring English until the last week. Phonetics needs daily, spaced practice.
8. Cramming the night before. Your memory consolidates while you sleep. A tired brain makes careless mistakes in exactly the questions you'd normally get right.

If you don't clear it#

Not clearing the qualifier on the first attempt is common, and it's recoverable. What you do next depends on which gate you missed.

You met the assignment cut-off but missed or failed the exam. You can re-attempt in the next qualifier exam within the same term by paying the partial re-attempt fee (₹2,000 for General/OBC). Your assignment scores carry over, so you don't redo the coursework. Use the gap to fix the course that let you down.

You didn't meet the assignment cut-off. You'll need to apply again for a later term as a new applicant, paying the full fee. Next time, treat assignments as the easy marks they should be.

Either way, change how you prepare, not just how hard. Look at your scores by course. If one course was below the minimum, it needs a different approach: more practice problems, a different explanation (our cheat notes, a study partner, office-hours sessions), or simply starting it first each day instead of last.

After you qualify#

Congratulations: you're in. A few things happen next.

• Registration. You register for Foundation courses for the next term, choosing how many (up to your course credit capacity). If you're also in college or working, start with two at most.
• Documents. Upload one of your Class 12 mark sheet, a degree certificate or your highest education certificate. Verification takes up to three weeks.
• Your qualifier score helps. It counts as Quiz 1 for the Foundation courses you take in the term immediately after the qualifier, which is a strong reason to register for Maths I, Stats I, CT and English I first.
• Validity. Your pass stays valid for three terms (six if you're still in school), so you don't have to start immediately if the timing is wrong, but don't let it lapse.

Our complete IITM BS guide explains what a full term looks like, how fees and waivers work, and how many courses to take.

Frequently asked questions#

What is the syllabus of the IITM BS qualifier exam?#

Weeks 1–4 of four courses: Mathematics for Data Science I (sets, relations, functions, straight lines, quadratics, polynomials), Statistics for Data Science I (types of data, categorical and numerical summaries, association), Computational Thinking (iteration, filtering, datasets, pseudocode, procedures) and English I (sounds, grammar, vocabulary, reading). Follow the official course content for your term.

What is the cut-off for the IITM BS qualifier?#

For the General category: at least 40% in every course and at least 50% average in the exam, after an assignment average of at least 40% in every course. OBC-NCL/EWS candidates need 35% and 45%; SC/ST/PwD candidates need 30% and 40%.

Is the qualifier exam online?#

No. The four weeks of content and assignments are online, but the qualifier exam itself is in person, at an exam centre in the city you chose.

How long is the qualifier exam?#

Four hours, covering all four courses.

Can I prepare for the qualifier before applying?#

Yes, and a week of light preparation helps. Revise Class 10 algebra, read our cheat notes once, and try a previous-year paper untimed to see the question style. Don't try to master everything early: the official content in the four weeks is what the exam follows.

How many hours a day do I need for the qualifier?#

Two to three hours a day, five or six days a week, is enough for most people. If you're working full-time, 1–1.5 hours on weekdays plus longer weekend sessions can work, as long as you never skip several days in a row.

What if I miss one weekly assignment?#

Gate 1 uses your best two of the first three assignments in each course, so one missed assignment can be covered by strong scores in the other two. Missing two of the first three in any course usually means you won't be eligible for the exam.

Can I take the qualifier while in Class 12?#

Yes. Students who have finished their Class 11 final exams can apply and take the qualifier. They can register for Foundation courses only after passing Class 12, and their pass is valid for six terms.

How many times can I attempt the qualifier?#

We couldn't find a limit on reapplying in the official admissions pages. Within a term, if you met the assignment cut-off, you can re-attempt the exam in the next qualifier exam at a partial fee. Otherwise you apply again in a later term.

Do qualifier assignments count if I submit late?#

No. Deadlines are fixed. Submit well before the deadline, in IST.

Which course is the hardest in the qualifier?#

It depends on your background, but CT surprises the most people because it looks easy and requires exact tracing. For students who left maths after Class 10, Maths I takes the most effort.

Is there negative marking in the qualifier?#

Check the instructions for your exam. Marking schemes are stated in the exam instructions, and you should read them on the day.

Does the qualifier score affect my Foundation grades?#

Yes, in a good way: it counts as Quiz 1 for the Foundation courses you take in the term immediately after the qualifier.

Can I change my exam city?#

You can change your exam city up to four weeks before the qualifier exam, according to the FAQ.

How should I divide time between the four courses?#

Divide by need, not equally. Give your weakest course the most time during preparation and your freshest hour during the exam. The per-course minimum means your weakest course decides your result.

Are previous-year qualifier papers available?#

Yes. BTechi's PYQ library has past qualifier papers with official answer keys, and our Qualifier Pack has full timed mocks. Use them topic by topic in weeks 1–3 and as full timed papers in week 4.

When does the January 2027 qualifier start?#

Applications for the January 2027 term close on 29 January 2027. See our January 2027 admission guide for the latest confirmed dates.