Learn

Sunday, 25 October

JEE Main Paper 1 · 108 shifts · 2017–2026

JEE Main previous year questions, chapter by chapter

Lumi tagged 9,285 questions from 108 official JEE Main shifts (2017–2026) to 81 chapters, 196 topics and 2,809 ideas. 53 chapters hold 80% of the questions, and 100% of the topic-tagged 2024–26 questions came from a topic asked in an earlier year.

Questions
9,285
44 dropped by NTA
Shifts analysed
108 / 174
held 2017–2026
Chapters for half
28 of 81
53 for 80%
Repeat ideas
1432
of 2,809 ideas asked

Key findings

What the analysed shifts show

From 108 of the 174 JEE Main shifts held from 2017 to 2026. The sections below show the detail.

  1. 53 of 81

    01
    28 → 50%53 → 80%chapters, most asked first

    chapters cover 80% of questions. The top 28 cover only half.

  2. 100% vs 79%

    02
    • Topic asked before100%
    • Exact idea asked before79%

    Topics repeat, and so do most exact ideas (2024–26 shifts, against earlier years).

  3. 95% vs 36%

    03
    • Topic in another shift95%
    • Exact idea in another shift36%

    Of a shift's topics and exact ideas, the share also in that year's other shifts (2024–26).

  4. 31% → 27%

    04
    31%27%20172026

    Easy questions, 2017–23 average against 2025–26.

  5. 32% → 42%

    05
    32%42%20172026

    Multi-step questions, 2017–23 against 2025–26.

  6. 5% → 9%

    06
    5%9%20172026

    Statement, assertion–reason and match-the-list questions, 2017–23 against 2025–26.

  7. 57 of 70

    07
    • Asked all 8 years
    • Missed a year or more

    chapters in the 2026 syllabus were asked in every one of the eight years analysed.

  8. +1.3 / -0.6

    08

    Biggest rise and fall in questions a shift (of 25 a subject), 2024–26 against 2017–23.

Of the 10,912 ideas Lumi's JEE taxonomy lists for these chapters, 2,809 (26%) were ever the main idea of a question. 11 chapters asked before 2024 are no longer in NTA's 2026 syllabus; the analysed 2024–26 shifts asked each of them at most 2 times.

Subjects

Physics, Chemistry and Maths

Each shift has 25 questions a subject (30 offered in 2021–24, of which 25 were answered; the 2017 paper had 30, all answered).

Questions by subject

Physics

33%

3,084 questions

Chemistry

33%

3,092 questions

Maths

33%

3,065 questions

33%33%33%
Physics29chapters asked30% numerical · 5% hard · 655 with a figure12 chapters hold halfChemistry30chapters asked30% numerical · 6% hard · 674 with a figure11 chapters hold halfMaths23chapters asked30% numerical · 26% hard · 22 with a figure8 chapters hold half
Class 11 or Class 12

  • Physics

    47%53%

  • Chemistry

    42%58%

  • Maths

    50%50%
  • Class 11
  • Class 12

Questions by the NCERT class of their chapter. Chemistry leans most on Class 12 (58%), Maths least (50%).

Easy questions by subject
  • Physics44%
  • Chemistry41%
  • Maths11%

Share rated easy across the 108 analysed shifts.

Hard Physics questions
0%20%40%’17’20’21’22’23’24’25’2676366

Share rated hard, by year.

Hard Chemistry questions
0%20%40%’17’20’21’22’23’24’25’261048810

Share rated hard, by year.

Hard Maths questions
0%20%40%’17’20’21’22’23’24’25’262419342630

Share rated hard, by year.

Chapter weightage

No short list of chapters

It takes 28 chapters to reach half the questions and 53 to reach 80%, in 108 of the 174 JEE Main shifts held from 2017 to 2026.

Concentration curve
0%25%50%75%100%12040608081chapters, ranked by questions in the analysed shifts28 chapters → 50%53 chapters → 80%

Share of all questions covered as chapters are added, biggest first. Dashed: every chapter equal.

The ten biggest chapters
running share
  1. 1
    Conic Sections296
    3%
  2. 2
    Three Dimensional Geometry224
    6%
  3. 3
    Integrals211
    8%
  4. 4
    Organic Chemistry: Some Basic Principles and Techniques207
    10%
  5. 5
    Complex Numbers and Quadratic Equations205
    12%
  6. 6
    Current Electricity193
    14%
  7. 7
    Coordination Compounds191
    17%
  8. 8
    Sequences and Series177
    18%
  9. 9
    System of Particles and Rotational Motion165
    20%
  10. 10
    Differential Equations161
    22%

Together they hold 22% of all questions.

Chapter map
70 chapters in the 2026 syllabus
  • Up 1+ a paper
  • Up 0.3 to 1
  • Steady
  • Down 0.3 to 1
  • Down 1+ a paper

Tile size is questions in the analysed shifts; colour is the change in questions a shift (of 25 a subject), 2024–26 against earlier years. 14 chapters are up, 5 down. Open a tile for the chapter.

Physics
all chapters
Chemistry
all chapters
Maths
all chapters

What the chapters are worth

Marks the biggest chapters held, shift by shift

Each question is worth 4 of 300 marks (75 questions answered). Adding chapters biggest first, this is how many marks they held in each of the 108 analysed shifts. The band runs from the leanest shift to the richest; the line is the middle one. It is history, not a promise for the next paper.

Historical marks coverage
075150225300120406070chapters, biggest first by average a shiftTop 10Top 20Top 30All live
  • Lowest to highest shift
  • Middle shift

Chapters in the 2026 syllabus, ranked by average questions a shift. Papers with 90 questions (2017 and 2021–24) are scaled to 75. Ranked with every analysed shift, so this is hindsight, not a test. In the middle shift, the top 29 held half the marks.

Read across
marks of 300
  • Top 10 chapters6447–92 across shiftsmiddle shift
  • Top 20 chapters11383–140 across shiftsmiddle shift
  • Top 30 chapters156123–195 across shiftsmiddle shift
  • All 70 chapters in the syllabus273266–300 across shiftsmiddle shift

Marks a shift held in these chapters, not marks anyone scored. Chapters dropped from the syllabus held part of older shifts, which is why some shifts fall short of 300 even with every chapter in the syllabus.

Average, range and weighted estimate

Questions a shift, by chapter

For the biggest chapters in the 2026 syllabus: the average and typical range of questions a shift (history) and a recency-weighted estimate (model).

Physics
of 25 a shift
Chemistry
of 25 a shift
Maths
of 25 a shift
  • Typical range (fact)
  • Average a shift (fact)
  • Weighted estimate (model)

Questions are counted per shift, out of the 25 each subject has, in the shifts Lumi analysed; papers that offered 30 a subject (2017, 2021–2024) are scaled to 25. Typical range: the middle six of the eight years (2017, 2020–2026), leaving out the single highest and lowest year. The average and the range are historical fact. The weighted estimate is a model: each year's share, with recent years counting more (half-life three years), and chapters NTA dropped from the syllabus set near zero. It is a guide to effort, not a forecast of the paper. In the backtest it did about as well as a plain all-time count; see how far to trust it.

Since 2024

What moved, and what left the syllabus

Questions a shift (of 25 a subject) in 2024–26 against 2017–23, for chapters in the 2026 syllabus. Chapters at about one question a shift move by chance too.

Asked more since 2024
Asked less since 2024
Surprise years and shifts

Chapters far above or below their own average in one shift, in questions of 25 a subject.

Difficulty

Fewer easy questions since 2025

Easy questions were 31% of a year in 2017–23 and 27% in 2025–26; hard went from 13% to 14%. Difficulty is Lumi's rating of each question, so read it as a trend.

Easy, medium and hard
0%50%100%29’1727’2034’2133’2234’2336’2427’2526’26
  • Easy
  • Medium
  • Hard

Number in each column: the easy share of that year's analysed shifts.

Questions that need a calculation or several steps
2025–260%50%100%’17’20’21’22’23’24’25’266360616672

63% a year in 2017–23, 69% in 2025–26.

How it asks

Formats and question style, year by year

The numerical-answer share is set by the exam pattern: none in 2017, 20% in 2020 and 2025–26, 33% offered in 2021–24. Everything else here is a choice the setters made.

Question formats
  • Single answer
  • Numerical answer
  • Statements
  • Match the list
  • Assertion–reason
  • Diagram
  • Other
Multi-step questions
0%30%60%’17’20’21’22’23’24’25’263628294144

Share of each year's questions that chain two or more results.

Statement, assertion–reason and match-the-list
0%10%20%’17’20’21’22’23’24’25’2606999

Share of each year's questions.

  • Single correct

    61%-37 since 2017
    20172026

  • Numerical value

    20%+20 since 2017
    20172026

  • Other

    6%+3 since 2017
    20172026

  • Two statements

    5%+5 since 2017
    20172026

  • Diagram based

    5%+5 since 2017
    20172026

  • Match the list

    3%+3 since 2017
    20172026

  • Assertion–reason

    1%+1 since 2017
    20172026
Year profile
YearShiftsNumericalMulti-stepCalculation or multi-stepRecallEasyHardWith a figureNew idea
2026920%44%72%6%26%15%13%18%
2025820%41%66%6%27%13%18%23%
20241833%30%63%10%36%12%14%23%
20231234%29%61%11%34%15%13%25%
20222433%32%65%10%33%14%13%36%
20212433%28%60%11%34%10%16%67%
20201220%37%67%8%27%11%18%95%
201710%36%63%8%29%14%19%–

Shares of the year's questions, each shift weighted equally. "New idea": questions on an idea not asked in any earlier year analysed (none for 2017, the first year).

Chapters that set the most numerical questions

Share of each chapter's questions with a typed numerical answer. Across all chapters: 30%.

Chapters that set the fewest

Live chapters with 20 or more questions.

Repeats

What comes back, and what does not

The share of each year's questions on a chapter, topic or exact idea already asked in an earlier year Lumi analysed.

Share of each year seen in an earlier year

2024–26 average

  • Same chapter100%
  • Same topic100%
  • Same exact idea79%
  • Same idea as the year before39%
0%25%50%75%100%’20’21’22’23’24’25’26100%100%82%
  • Same chapter
  • Same topic
  • Same exact idea

2026: chapter 100.0%, topic 100.0%, exact idea 81.8%.

New ideas in each year
2024–260%50%100%’17’20’21’22’23’24’25’2695362318

Tested an idea no earlier year had asked. Early years run high: less history came before them. 2024–26 average: 22%.

Same idea as the year before
0%30%60%’17’20’21’22’23’24’25’265553929

Share of a year's questions repeating an exact idea from the analysed year before it (2020 is set against 2017). Years with more analysed shifts have more to match. 2024–26 average: 39%.

Ideas asked most
all 1432 repeats
Ideas never asked, by chapter

Of the ideas Lumi's taxonomy lists for a chapter, the share never the main idea of a question. Chapters with 40 or more listed ideas.

The exact-idea share rises from 4.6% in 2020 to 81.8% in 2026 as the history grows: 2020 is matched against 1 earlier analysed shift, 2026 against 99. 196 topics and 2,809 ideas are counted; questions with no close idea (1,535) are left out of the topic and idea lines.

Shift to shift

Do later shifts repeat earlier shifts?

JEE Main runs many shifts a session, each a different paper. Here is how much of each 2024–26 shift was already in the same year's other shifts.

Share of a shift already in the year's other shifts

  • 2024

    • Topics99 %
    • Exact ideas53 %

  • 2025

    • Topics92 %
    • Exact ideas27 %

  • 2026

    • Topics94 %
    • Exact ideas28 %

Averages over the analysed shifts of each year.

How alike two shifts are
  • Same year · 901 pairs0.38
  • Different years · 4877 pairs0.3
  • Same year, 2024–260.55
  • Different years, 2024–260.46

Rank correlation of the chapter mix between two shifts (1 = same order). Over every pair of analysed shifts.

For topics, nearly always: 92%–99% of a shift's topics also came in another shift that year. For exact ideas, 27%–53%. Shifts in the same year look a little more alike than shifts in different years. The 2025 and 2026 shifts here are all from the April session, so for those years this compares shifts within one session. The within-year test checks whether adding the year's other shifts to the history helps.

Year by year

Every year at a glance

Bar: easy, medium and hard. Open a year for each shift and every question.

High-yield topics

Topics asked in at least seven of eight years

124 topics came in at least 7 of the 8 years analysed. The biggest in each subject, with questions a shift (of 25 a subject): what history shows, and what a recency-weighted model estimates. Dots: the years each was asked.

Physics

Chemistry

Maths

  • Typical range (fact)
  • Average a shift (fact)
  • Weighted estimate (model)

Questions are counted per shift, out of the 25 each subject has, in the shifts Lumi analysed; papers that offered 30 a subject (2017, 2021–2024) are scaled to 25. Typical range: the middle six of the eight years (2017, 2020–2026), leaving out the single highest and lowest year. The average and the range are historical fact. The weighted estimate is a model: each year's share, with recent years counting more (half-life three years), and chapters NTA dropped from the syllabus set near zero. It is a guide to effort, not a forecast of the paper.

Chapter-wise weightage

Every chapter, every year

For each chapter: the average and typical range of questions a shift, of 25 a subject (history), and a recency-weighted estimate (model). Switch to the heatmap for questions a shift in each year, so a year with more analysed shifts (1 to 24 a year) does not look bigger.

70 chapters · 8626 questions asked since 2017
  • Conic SectionsMust do2–3.1Mathematics · 11 · 296 in 108 shifts+0.6avg 2.4 · est 2.7Chapter page
  • Organic Chemistry: Some Basic Principles and TechniquesMust do1–2.7Chemistry · 11 · 207 in 108 shifts+1.3avg 1.6 · est 2.1Chapter page
  • Three Dimensional GeometryMust do1.7–2Mathematics · 12 · 224 in 108 shifts · Part dropped0.1avg 1.8 · est 1.9Chapter page
  • Coordination CompoundsMust do1.1–2.1Chemistry · 12 · 191 in 108 shifts+0.4avg 1.6 · est 1.9Chapter page
  • Complex Numbers and Quadratic EquationsMust do1.4–1.9Mathematics · 11 · 205 in 108 shifts+0.3avg 1.6 · est 1.7Chapter page
  • IntegralsMust do1.6–1.9Mathematics · 12 · 211 in 108 shifts0.0avg 1.7 · est 1.7Chapter page
  • Sequences and SeriesMust do1.4–1.8Mathematics · 11 · 177 in 108 shifts0.1avg 1.5 · est 1.6Chapter page
  • Haloalkanes and HaloarenesMust do0.8–1.8Chemistry · 12 · 141 in 108 shifts0.0avg 1.5 · est 1.5Chapter page
  • Ray Optics and Optical InstrumentsMust do0.9–1.7Physics · 12 · 148 in 108 shifts+0.5avg 1.3 · est 1.4Chapter page
  • Aldehydes, Ketones and Carboxylic AcidsMust do1.1–1.5Chemistry · 12 · 159 in 108 shifts · Part dropped-0.1avg 1.4 · est 1.4Chapter page
  • The d- and f-Block ElementsMust do0.8–1.4Chemistry · 12 · 153 in 108 shifts+0.6avg 1.2 · est 1.4Chapter page
  • Current ElectricityMust do1–1.9Physics · 12 · 193 in 108 shifts-0.5avg 1.6 · est 1.4Chapter page
  • AminesMust do0.7–1.6Chemistry · 12 · 152 in 108 shifts+0.4avg 1.1 · est 1.4Chapter page
  • Units and MeasurementsMust do0.8–1.7Physics · 11 · 148 in 108 shifts+0.7avg 1.2 · est 1.4Chapter page
  • System of Particles and Rotational MotionMust do1.1–1.8Physics · 11 · 165 in 108 shifts-0.1avg 1.4 · est 1.4Chapter page
  • DeterminantsMust do1.1–1.5Mathematics · 12 · 160 in 108 shifts-0.3avg 1.5 · est 1.3Chapter page
  • Relations and FunctionsMust do1.1–1.5Mathematics · 11 · 158 in 108 shifts0.3avg 1.2 · est 1.3Chapter page
  • Moving Charges and MagnetismMust do0.9–1.5Physics · 12 · 155 in 108 shifts0.2avg 1.2 · est 1.3Chapter page
  • EquilibriumMust do0.8–1.2Chemistry · 11 · 127 in 108 shifts0.2avg 1.1 · est 1.3Chapter page
  • SolutionsShould do0.9–1.2Chemistry · 12 · 124 in 108 shifts+0.3avg 1.1 · est 1.3Chapter page
  • Differential EquationsShould do1–1.5Mathematics · 12 · 161 in 108 shifts0.2avg 1.2 · est 1.3Chapter page
  • The p-Block ElementsMust do0.8–1.5Chemistry · 12 · 146 in 108 shifts · Part dropped-0.2avg 1.2 · est 1.3Chapter page
  • BiomoleculesShould do0.8–1.4Chemistry · 12 · 122 in 108 shifts+0.3avg 1.0 · est 1.2Chapter page
  • Chemical Bonding and Molecular StructureMust do0.9–1.1Chemistry · 11 · 142 in 108 shifts+0.3avg 1.1 · est 1.2Chapter page
  • Thermodynamics (Chemistry)Should do0.7–1.4Chemistry · 11 · 112 in 108 shifts0.2avg 1.1 · est 1.2Chapter page

Right: typical range of questions a shift, and the change since 2024. Avg: average a shift (fact). Est: weighted estimate (model).

  • Typical range (fact)
  • Average a shift (fact)
  • Weighted estimate (model)

Questions are counted per shift, out of the 25 each subject has, in the shifts Lumi analysed; papers that offered 30 a subject (2017, 2021–2024) are scaled to 25. Typical range: the middle six of the eight years (2017, 2020–2026), leaving out the single highest and lowest year. The average and the range are historical fact. The weighted estimate is a model: each year's share, with recent years counting more (half-life three years), and chapters NTA dropped from the syllabus set near zero. It is a guide to effort, not a forecast of the paper.

How far to trust it

A map of the paper, not a forecast

We tested the estimate on the 2022–2026 shifts, each time using only the years before. Counting past shifts beats picking at random by a wide margin: the top 20 topics held 23.5% of a shift, against 10.2% for a random 20. Against a plain all-time count, weighting recent years more comes out level with it (+0.7 points, under one question a shift) at topic level and level with it (+0.9 points, under one question a shift) at chapter level. It never names the questions.

The full backtest, year by year
  • Top 20 topicsTie · +0.5 Q
    • Weighted estimate24%
    • All-time count23%
    • Last year only24%
    • Random pick10%
  • Top 30 chaptersTie · +0.7 Q
    • Weighted estimate53%
    • All-time count52%
    • Last year only52%
    • Random pick37%

Share of the next year’s questions held by each method’s top picks, averaged over 5 years of shifts.

Explore

JEE Main past papers, every way round

What this covers

108 of 174 shifts

It covers 108 of the 174 JEE Main shifts held from 2017 to 2026 (9,285 Paper 1 questions), not every shift. Lumi analyses a shift when NTA's full paper and final answer key are both available. Every shift held, and its status:

2026

9 of 19 analysed

  • 21 Jan S1: Paper not published
  • 21 Jan S2: Paper not published
  • 22 Jan S1: Paper not published
  • 22 Jan S2: Paper not published
  • 23 Jan S1: Paper not published
  • 23 Jan S2: Paper not published
  • 24 Jan S1: Paper not published
  • 24 Jan S2: Paper not published
  • 28 Jan S1: Paper not published
  • 28 Jan S2: Paper not published
  • 2 Apr S1
  • 2 Apr S2
  • 4 Apr S1
  • 4 Apr S2
  • 5 Apr S1
  • 5 Apr S2
  • 6 Apr S1
  • 6 Apr S2
  • 8 Apr S2
2025

8 of 19 analysed

  • 22 Jan S1: Not analysed yet
  • 22 Jan S2: Not analysed yet
  • 23 Jan S1: Not analysed yet
  • 23 Jan S2: Not analysed yet
  • 24 Jan S1: Paper not published
  • 24 Jan S2: Not analysed yet
  • 28 Jan S1: Not analysed yet
  • 28 Jan S2: Not analysed yet
  • 29 Jan S1: Paper not published
  • 29 Jan S2: Paper not published
  • 2 Apr S1
  • 2 Apr S2
  • 3 Apr S1
  • 3 Apr S2
  • 4 Apr S1: Not analysed yet
  • 4 Apr S2
  • 7 Apr S1
  • 7 Apr S2
  • 8 Apr S2
2024

18 of 20 analysed

2023

12 of 24 analysed

2022

24 of 24 analysed

2021

24 of 30 analysed

2020

12 of 16 analysed

2019

0 of 16 analysed

  • 9 Jan S1: No final key
  • 9 Jan S2: Paper not published
  • 10 Jan S1: Paper not published
  • 10 Jan S2: Paper not published
  • 11 Jan S1: Paper not published
  • 11 Jan S2: Paper not published
  • 12 Jan S1: Paper not published
  • 12 Jan S2: Paper not published
  • 8 Apr S1: Paper not published
  • 8 Apr S2: Paper not published
  • 9 Apr S1: Paper not published
  • 9 Apr S2: Paper not published
  • 10 Apr S1: Paper not published
  • 10 Apr S2: Paper not published
  • 12 Apr S1: Paper not published
  • 12 Apr S2: Paper not published
2018

0 of 3 analysed

  • 8 Apr S1: Paper not published
  • 15 Apr S1: Paper not published
  • 16 Apr S1: Paper not published
2017

1 of 3 analysed

  • 2 Apr offline
  • 8 Apr online: Paper not published
  • 9 Apr online: Paper not published

Blue: analysed. Dashed: not in this analysis, because NTA did not publish the paper (52 shifts) or it has not been analysed yet. 44 questions NTA dropped are shown on their shift but left out of every count. No answer key needed an erratum.

Questions

About this analysis

Can anyone predict the exact JEE Main questions?
No. Topics come back, and so do most exact ideas, but in 2024–26 about 22% of each year's questions tested an idea no earlier analysed year had asked. Past questions show what was asked, not what the next shift will ask.
Which topics come almost every year?
124 topics were asked in at least 7 of the 8 years analysed (Kinetic Theory of Gases and Kinematics in 1D; d-Block Elements and Crystal Field Theory; Permutations and Combinations and AP and GP, among others). That is a pattern in past shifts, not a promise for the next one.
How many JEE Main shifts does this cover?
108 of the 174 JEE Main Paper 1 shifts held from 2017 to 2026: 2026: 9 of 19 · 2025: 8 of 19 · 2024: 18 of 20 · 2023: 12 of 24 · 2022: 24 of 24 · 2021: 24 of 30 · 2020: 12 of 16 · 2019: 0 of 16 · 2018: 0 of 3 · 2017: 1 of 3. Lumi analyses a shift when the full official paper and its final answer key are both available. No 2018 or 2019 shift is in yet.
Which JEE Main chapters carry the most questions?
In the analysed shifts: Conic Sections (296), Three Dimensional Geometry (224), Integrals (211), Organic Chemistry: Some Basic Principles and Techniques (207), Complex Numbers and Quadratic Equations (205). 28 of 81 chapters hold half of all questions and 53 hold 80%.
Do JEE Main questions repeat?
Topics do, and so do most exact ideas. In 2024–26, 100.0% of questions came from a topic already asked in an earlier year. At the level of the exact idea, 76.8% (2024) to 81.8% (2026) had been asked before. Each year is matched against a longer history: 2024 against 73 earlier analysed shifts, 2026 against 99.
Do later shifts repeat questions from earlier shifts of the same session?
Topics, nearly always; exact ideas, much less. Across 2024–26, 92%–99% of a shift's topics also came in that year's other shifts, against 27%–53% of its exact ideas.
What is the weighted estimate?
A weighted average of each chapter's share of past shifts, with recent years counting more, scaled to 25 questions a subject. Tested on 2022–2026, its top 30 chapters held 53% of a shift's questions, within a question a shift of a plain all-time count (52%) and against 37% for a random 30. It is a guide to effort, not a forecast of the paper.

Updated 2026-10-24. By Lumi.