JEE Main 2027 Marks vs Percentile: How NTA Converts Scores
NTA converts marks into a percentile by ranking you inside your own session: percentile = 100 x (candidates with a score at or below yours) / (all candidates in that session). That is why no fixed marks-to-percentile table exists. This page explains the method, shows what 2026 looked like and how to estimate without fooling yourself.
Last verified: 15 November 2026
Key facts at a glance
- Formula
- 100 x (candidates scoring <= you) / (total candidates in that session)
- Is it fixed?
- No. It changes with each session, shift and year
- 2026 Session 1 appeared
- About 13.05 lakh (reported: 13,04,653)
- 2026 General JEE Advanced cut-off
- 93.4123549 percentile (NTA, 20 April 2026)
- 2027 Session 1
- 20, 21, 22, 23, 24 and 28 January 2027
- Status (15 Nov 2026)
- No 2027 marks-percentile data exists; 2026 shown for reference only
Status: what exists and what does not
How NTA converts raw marks into a percentile
The NTA score in JEE Main is a percentile, not a percentage. For a given session, NTA lists all candidates by their score and gives each candidate the share of the session's candidates who scored equal to or less than them. In formula form: percentile = 100 x (number of candidates with score less than or equal to yours) / (total number of candidates in that session).
A candidate at the 99th percentile has therefore beaten or tied with 99 per cent of the people who appeared in that session. With about 13.05 lakh candidates in Session 1 of 2026, each percentile point near the top stands for roughly 13,000 candidates, so a single mark can shift your percentile by a visible amount in the dense middle of the distribution and by a tiny amount at the very top.
When a session is spread over several shifts, the raw scores of different shifts are not directly comparable, because the papers differ. Per Appendix IV, the highest raw score in every shift becomes 100 percentile, the lowest gets a small value that depends on the number of candidates in that shift, and the shift percentiles are merged and called the NTA scores, which then drive the merit list. Our page on the normalization process explains this in more detail.
The bulletin works an example for a shift of 41,326 candidates: a candidate with 121 raw marks and 37,244 candidates at or below that score gets (37,244 / 41,326) x 100 = 90.1224411 percentile, while 41 raw marks with 20,851 at or below gives 50.4549194. Percentiles are carried to 7 decimal places to avoid bunching and reduce ties.
Eligibility cut-offs follow the same logic: if the cut-off marks correspond to different percentiles in different shifts, the lowest percentile becomes the cut-off for that category in all shifts (bulletin example: 78 in one shift and 79 in another makes 78 the cut-off). Paper 1 merit ties are broken by NTA score in Mathematics, then Physics, then Chemistry, then lower proportion of incorrect to correct answers (bulletin 2.4).
Why marks to percentile is not fixed
- Paper difficulty: a harder shift lowers every score, so the marks needed for 99 percentile fall. In 2026, one coaching analysis estimates the 99th percentile mark ranged from about 154 in the toughest Session 1 shift to about 188 in the easiest.
- Candidate pool: a stronger pool pushes the same marks to a lower percentile. Session 2 attracts different candidates than Session 1 (some skip it, some are repeaters improving a score).
- Distribution shape: marks are dense between 40 and 100 and thin at the top, so a 10-mark change moves the percentile a lot in the middle and very little at the top.
- Negative marking: Paper 1 awards +4 for a correct answer and -1 for a wrong one in both the MCQ and numerical-value sections (bulletin 2.4), so the spread depends on how many guesses go wrong.
- Year to year: attendance, difficulty and the share of repeaters change. The average 99th-percentile mark in one analysis rose from about 170 in 2025 to about 174 in 2026, which shows a drift even when the paper style is the same.
2026 reference points (labelled 2026, not 2027)
The tables below are what the 2026 cycle looked like according to published analyses. They are shift-wise estimates by private organisations, not NTA data. Where two sources overlap they broadly agree but are not identical, which is itself a reminder not to use a single number.
| 2026 raw score band | Approx. percentile (2026 Session 1) | Approx. rank (2026) | Source |
|---|---|---|---|
| 300 to 291 | 100 to 99.99989 | 1 to 15 | PW result analysis |
| 290 to 281 | 99.999 to 99.997 | 16 to 36 | PW result analysis |
| 270 to 261 | about 99.990 | 101 to 160 | PW result analysis |
| 200 to 191 | about 99.6 to 99.7 | 3,804 to 5,320 | PW result analysis |
| 160 to 151 | about 98.7 to 99.0 | 13,164 to 17,290 | PW result analysis |
| 100 to 91 | about 94.3 to 95.0 | 61,652 to 79,298 | PW result analysis |
Approximate figures from a coaching site's 2026 analysis (published after the Session 1 result). They are not published by NTA. PW itself calls them approximate.
99th percentile marks by shift in 2026 (estimates)
Read the spread, not the individual cells. Inside one session the 99th percentile mark moved by more than 30 marks from the hardest to the easiest shift. A candidate who plans for exactly 160 marks and lands in an easier shift could miss the 99th percentile by a visible margin.
| Session 1 shift (2026) | 99th percentile marks | Session 2 shift (2026) | 99th percentile marks |
|---|---|---|---|
| 21 Jan, shift 1 | 158 | 2 Apr, shift 1 | 175 |
| 21 Jan, shift 2 | 174 | 2 Apr, shift 2 | 165 |
| 22 Jan, shift 1 | 186 | 4 Apr, shift 1 | 175 |
| 22 Jan, shift 2 | 154 | 4 Apr, shift 2 | 174 |
| 23 Jan, shift 1 | 166 | 5 Apr, shift 1 | 190 |
| 23 Jan, shift 2 | 158 | 5 Apr, shift 2 | 175 |
| 24 Jan, shift 1 | 174 | 6 Apr, shift 1 | 160 |
| 24 Jan, shift 2 | 178 | 6 Apr, shift 2 | 183 |
| 28 Jan, shift 1 | 188 | - | - |
| 28 Jan, shift 2 | 180 | - | - |
Competishun's own shift-wise estimates for 2026, not official NTA data. The same page gives roughly 230 marks for the 99.9th percentile and roughly 113 (January) and 124 (April) marks for the 95th percentile on average. Its summary range and its band table differ slightly.
The only official 2026 percentile anchors: JEE Advanced cut-offs
NTA published category-wise cut-off percentiles on 20 April 2026 for JEE Advanced 2026 qualification, in the roughly 2.5 lakh candidate pool. These are percentiles, not marks, and they come from the combined result of both sessions.
| Category (2026) | Cut-off percentile for JEE Advanced 2026 |
|---|---|
| General | 93.4123549 |
| Gen-EWS | 82.4164528 |
| OBC-NCL | 80.9232583 |
| SC | 63.9172792 |
| ST | 52.0174712 |
2026 values as published by NTA on 20 April 2026. 2027 cut-offs will be published only after both 2027 sessions.
How to estimate your percentile safely after the exam
- 1
Compute your raw marks from the response sheet
After each session NTA releases response sheets and a provisional answer key. Count +4 for each correct answer and -1 for each wrong answer in both Section A (MCQs) and Section B (numerical-value questions); unanswered questions score 0 (bulletin 2.4). See the Session 1 answer key page for the process.
- 2
Locate your shift and its difficulty
Read the shift-wise analysis for your date and shift. Note whether your shift was reported as easy, moderate or difficult by several sources, not one.
- 3
Use a band, not a point
Take the nearest 2026 reference rows and widen them by about 10 to 15 marks on each side. If your shift was easier than the 2026 average, move your expected percentile down; if harder, up.
- 4
Cross-check with two or more sources
If two independent coaching sites give similar bands for your shift, the estimate is more reliable. If they differ widely, assume the lower percentile for planning.
- 5
Plan on the lower end
Use the lower end of your band for deciding whether you need Session 2 and for first college lists. Replace the estimate with the real percentile the moment the scorecard is out.
Common mistakes when using marks vs percentile
- Reading a marks-to-percentile table as an official NTA conversion. NTA does not publish one.
- Using a 2025 or 2026 table for 2027 and treating it as a target. Difficulty changes every year.
- Adding or averaging your Session 1 and Session 2 marks. Only the better percentile is used.
- Ignoring the shift. The same marks in two shifts can differ by several percentile points.
- Treating the percentile as percentage of marks. A 90 percentile does not mean 90 per cent marks.
- Quoting a rank from a percentile without knowing the number of candidates. See the percentile to rank page.
- Calculating marks from a wrong answer key. The provisional key can change after challenges; recheck after the final key.
Worked example with 2026 numbers
Suppose a candidate in 2026 Session 1 estimated their score at about 160. The PW analysis lists 160 to 151 marks at roughly the 98.7 to 99.0 percentile. The Competishun shift table says the 99th percentile mark was about 158 in the 21 January morning shift but about 186 in the 22 January morning shift. So the same 160 marks might be just above 99 in one shift and nearer 98 in another.
The safe reading is therefore a band of about 98 to 99.2, and the candidate should plan with the lower end. If their aim was the General cut-off for JEE Advanced (93.41 in 2026), they were comfortably above it on any reading. If their aim was a high-ranking NIT branch, the extra decimal places of percentile matter and they should use Session 2.
What to watch before using any 2027 number
- Shift-wise analyses after each January 2027 shift, from at least two sources.
- The NTA result and scorecard, which is the only exact conversion for your own marks.
- The Session 2 timetable (2, 3, 4, 5, 6 and 7 April 2027), if you plan to attempt both sessions; its form opens on 5 January 2027.
- The pattern in bulletin 2.4: 75 questions and 300 marks for Paper 1, with 25 questions per subject.
Sources
JEE Main 2027: complete guide
Dates, eligibility, pattern, syllabus, result and counselling in one place.
Frequently Asked Questions
Answers on marks vs percentile, from the CSAB 2026 brochure
1How does NTA convert JEE Main marks into a percentile?
2Is there a fixed marks-to-percentile table for JEE Main 2027?
3How many marks are needed for 99 percentile in JEE Main?
4Why does the same score give different percentiles in different shifts?
5Does a higher percentile always mean more marks?
6Which percentile counts if I appear in both sessions?
7Can I trust the marks-vs-percentile tables on coaching sites?
8Is the 2026 table valid for JEE Main 2027?
9What was the JEE Advanced qualifying percentile in 2026?
10Where do I see my percentile after the exam?
Know Your College
Before Counselling
Predict your college using JoSAA 2026 cutoff data across IITs, NITs, IIITs and GFTIs.
Related Guides
Source
The percentile formula and normalization steps are from Appendix IV of NTA's JEE Main 2027 Information Bulletin; the 2026 category cut-offs were published by NTA. Marks reference points are from coaching-institute analyses (Competishun and PW), not from NTA. Last checked 15 November 2026.
