Independent tool — not affiliated with NTA, JoSAA, CBSE, or any government body. Predictions are estimates only. Always verify your actual results at jeemain.nta.nic.in and josaa.nic.in
Open Data

How We Predict Your Rank

We don't claim a magic accuracy number. Instead, here's every data point we used, the exact formula, and exactly where this tool will be wrong.

⚡Most predictors show you a number. We show you how we got there — the raw anchor data, the interpolation math, the error zones, and the honest limitations.
Section 1

The 8 Calibration Anchor Points

These are the official NTA 2025 marks-to-percentile figures from published press releases and scorecard distributions. Our entire prediction model is anchored to these 8 real values.

96%97%98%99%99.5%99.9%100%100140180220260MARKS (out of 300)PERCENTILE100→96.060%120→96.980%140→98.730%160→99.240%180→99.570%200→99.780%240→99.956%280→99.996%NTA 2025 Official Anchor PointInterpolation model±0.5% band

These 8 points are taken directly from NTA official press releases and scorecard distributions. Our model linearly interpolates between them. The shaded band shows ±0.5 percentile — the typical interpolation uncertainty between adjacent anchor points.

MarksNTA Official %ileOur PredictionDifferenceSource
28099.996%99.996%0.000%NTA 2025 press release
24099.956%99.956%0.000%NTA 2025 press release
20099.780%99.780%0.000%NTA 2025 press release
18099.570%99.570%0.000%NTA 2025 press release
16099.240%99.240%0.000%NTA 2025 press release
14098.730%98.730%0.000%NTA 2025 press release
12096.980%96.980%0.000%NTA 2025 press release
10096.060%96.060%0.000%NTA 2025 press release

Why 0% difference at all anchor points? Anchor points have 0% error by definition — they are the source data our model is calibrated to. Error enters only when interpolating between two adjacent anchor points.

Section 2

The Honest Error Map

Prediction reliability varies sharply across the marks range. Here's exactly how confident you should be depending on your score.

250 – 300 marksHIGH CONFIDENCE±0.5 percentile

Very few candidates score here. NTA publishes exact distribution figures for top scorers. Our model is essentially exact between 250–280 (two anchor points only 30 marks apart).

160 – 250 marksGOOD CONFIDENCE±1–2 percentile

The competitive engineering cutoff zone. 4 anchor points (160, 180, 200, 240 marks) provide dense coverage. Interpolation error is low but shift normalization can add ±1 percentile.

90 – 160 marksMODERATE±2–4 percentile

The highest density of students — roughly 40% of all 1.55 crore candidates fall here. Normalization variance is highest and our interpolation spans a 40-mark gap (100→140). Use predictions as a broad guide.

0 – 90 marksLOW CONFIDENCE±5+ percentile

NTA rarely publishes data below 90 marks. Our model extrapolates from the 90-100 gap using historical patterns. Predictions here are directional estimates only — do not use for planning.

Confidence level by marks range
250–300
95%
160–250
78%
90–160
55%
0–90
28%
Section 3

The Exact Formula

For any marks value between two anchor points, we use linear interpolation on the percentile scale. No machine learning, no black boxes — just a transparent formula you can verify yourself.

// MARKS → PERCENTILE (linear interpolation between anchors)
Given: marks M between anchor points (M₁, P₁) and (M₂, P₂)
Where P₁ > P₂ (higher marks = higher percentile, M₁ > M₂)
Percentile = P₂ + (( M − M₂ ) / ( M₁ − M₂ )) × ( P₁ − P₂ )
// PERCENTILE → AIR (All India Rank)
Rank = ceil( (( 100 − Percentile ) / 100 ) × Total_Candidates )
Where Total_Candidates = 1,550,000 (JEE Main 2026 estimated unique appeared)
Worked Example: 170 marks
Nearest anchors: (M₁=180, P₁=99.570%) and (M₂=160, P₂=99.240%)
Percentile = 99.240 + ((170 − 160) / (180 − 160)) × (99.570 − 99.240)
Percentile = 99.240 + (10/20) × 0.330 = 99.405%
Rank = ceil(((100 − 99.405) / 100) × 1,550,000) = ~9,225
Shift Normalization Adjustment

NTA normalizes scores across shifts using equi-percentile normalization. A student scoring 180 marks in a tough shift and a student scoring 195 marks in an easy shift may receive the same final percentile. We account for this with a ±1.5 percentile adjustment window for shift difficulty in the detailed calculator. This is why our rank ranges are shown as a band, not a single number.

Section 4

Where This Will Be Wrong

Honest disclaimer: No predictor is perfect. This section exists because we believe you deserve to know the specific failure modes before you use a prediction for planning. Read this carefully if your rank estimate is in a high-stakes zone.
01
Shift variance we cannot see

We do not have shift-specific data for 2026 yet. Our model uses 2025 patterns. JEE Main difficulty varies 3–5 percentile points per shift — a student scoring 170 in a hard shift might get the same percentile as someone scoring 185 in an easy shift.

02
The 95–99 percentile danger zone

Approximately 8 lakh students compete for ranks 15,000–1,50,000. A ±1 percentile error here equals ±15,000 rank swing. This is where our predictions are least reliable despite appearing small in percentile terms.

03
Category seat matrix changes

NTA occasionally changes the seat matrix. Our college predictor uses JoSAA 2025 cutoffs which may shift ±5–10% in JoSAA 2026. New institutes added, seats reduced, or reservation changes can move cutoffs unpredictably.

04
Numerical Value Question (NVQ) section variability

The NVQ section scoring is harder to normalize. Students who attempt 5 vs 3 NVQs see different effective marks distributions. Our model does not distinguish NVQ attempt rates across shifts.

05
No claim on JEE Advanced

JEE Advanced is a completely different exam with different question types, marking schemes, and normalization. Our Advanced predictor is rough directional estimation only — treat it as a planning aid, not a prediction. Error margin: ±20–30% on AIR.

Section 5

Data Sources

Every data point we use has a named, verifiable source. No scraped estimates, no aggregated approximations without attribution.

DataSourceLast UpdatedCoverage
Marks → Percentile curveNTA official press releasesJEE Main 20258 anchor points
JoSAA cutoffsjosaa.nic.in official allotmentJoSAA 2025 Round 6128 colleges
JEE Advanced AIR tableJEE Advanced official statistics2024Top 20,000 CRL
Historical cutoffsjosaa.nic.in2022–2024Major IITs / NITs
Shift-wise difficulty dataNTA scorecards + Careers360JEE Main 2026 S18 Jan shifts
Total candidatesNTA registration data2026 est.~1.55 crore
Last reviewed: April 2026If you find an error in our data, open an issue on our GitHub. We update this page within 48 hours of official NTA data releases. For official results, always visit jeemain.nta.nic.in.

Now that you know exactly how we predict — try it yourself.

Predict My RankFull Methodology