✓
2
3
NEET-PG is marked +4 for a correct answer and −1 for a wrong one.
Built from publicly published NBEMS result data. No individual candidate records are stored or shown. Methodology
Official marks vs rank — previous years
The published NEET-PG record: the rank each marks range actually secured (out of 800 marks).
| Marks | Rank range | Candidates |
|---|---|---|
| 700–724 | 1 – 4 | 4 |
| 675–699 | 5 – 43 | 39 |
| 650–674 | 44 – 300 | 257 |
| 625–649 | 301 – 1,129 | 829 |
| 600–624 | 1,130 – 3,090 | 1,961 |
| 575–599 | 3,091 – 6,517 | 3,427 |
| 550–574 | 6,518 – 11,310 | 4,793 |
| 525–549 | 11,311 – 17,547 | 6,237 |
| 500–524 | 17,548 – 24,966 | 7,418 |
| 475–499 | 24,967 – 32,813 | 7,847 |
| 450–474 | 32,814 – 41,589 | 8,776 |
| 425–449 | 41,590 – 50,855 | 9,266 |
| 400–424 | 50,856 – 60,381 | 9,526 |
| 375–399 | 60,382 – 70,360 | 9,979 |
| 350–374 | 70,361 – 80,825 | 10,463 |
| 325–349 | 80,826 – 91,677 | 10,852 |
| 300–324 | 91,678 – 1,03,605 | 11,927 |
| 275–299 | 1,03,606 – 1,16,150 | 12,543 |
| 250–274 | 1,16,151 – 1,29,853 | 13,702 |
| 225–249 | 1,29,854 – 1,44,698 | 14,844 |
| 200–224 | 1,44,699 – 1,60,146 | 15,448 |
| 175–199 | 1,60,147 – 1,75,919 | 15,772 |
| 150–174 | 1,75,920 – 1,90,772 | 14,850 |
| 125–149 | 1,90,773 – 2,04,399 | 13,626 |
| 100–124 | 2,04,400 – 2,15,363 | 10,963 |
| 75–99 | 2,15,364 – 2,22,966 | 7,602 |
| 50–74 | 2,22,967 – 2,27,379 | 4,410 |
| 25–49 | 2,27,380 – 2,29,435 | 2,056 |
| 0–24 | 2,29,436 – 2,30,019 | 584 |
| -25–-1 | 2,30,020 – 2,30,110 | 91 |
| -50–-26 | 2,30,111 – 2,30,114 | 4 |
2,30,096 candidates appeared in 2025.
| Marks | Rank range | Candidates |
|---|---|---|
| 725–749 | 1 – 1 | 1 |
| 700–724 | 2 – 20 | 19 |
| 675–699 | 21 – 91 | 71 |
| 650–674 | 92 – 412 | 321 |
| 625–649 | 413 – 1,105 | 693 |
| 600–624 | 1,106 – 2,586 | 1,481 |
| 575–599 | 2,587 – 4,837 | 2,251 |
| 550–574 | 4,838 – 8,347 | 3,510 |
| 525–549 | 8,348 – 12,945 | 4,598 |
| 500–524 | 12,946 – 18,467 | 5,522 |
| 475–499 | 18,468 – 25,104 | 6,637 |
| 450–474 | 25,105 – 32,713 | 7,608 |
| 425–449 | 32,714 – 41,006 | 8,293 |
| 400–424 | 41,007 – 50,221 | 9,215 |
| 375–399 | 50,222 – 60,274 | 10,053 |
| 350–374 | 60,275 – 71,142 | 10,868 |
| 325–349 | 71,143 – 82,974 | 11,832 |
| 300–324 | 82,975 – 96,094 | 13,120 |
| 275–299 | 96,095 – 1,10,162 | 14,068 |
| 250–274 | 1,10,163 – 1,25,075 | 14,913 |
| 225–249 | 1,25,076 – 1,40,379 | 15,304 |
| 200–224 | 1,40,380 – 1,55,340 | 14,961 |
| 175–199 | 1,55,341 – 1,68,903 | 13,563 |
| 150–174 | 1,68,904 – 1,80,299 | 11,396 |
| 125–149 | 1,80,300 – 1,88,795 | 8,496 |
| 100–124 | 1,88,796 – 1,94,619 | 5,824 |
| 75–99 | 1,94,620 – 1,98,095 | 3,476 |
| 50–74 | 1,98,096 – 1,99,777 | 1,682 |
| 25–49 | 1,99,778 – 2,00,343 | 566 |
| 0–24 | 2,00,344 – 2,00,504 | 161 |
| -25–-1 | 2,00,505 – 2,00,516 | 12 |
| -50–-26 | 2,00,517 – 2,00,517 | 1 |
2,00,516 candidates appeared in 2023.
| Marks | Rank range | Candidates |
|---|---|---|
| 700–724 | 1 – 1 | 1 |
| 675–699 | 2 – 26 | 25 |
| 650–674 | 27 – 142 | 116 |
| 625–649 | 143 – 463 | 321 |
| 600–624 | 464 – 1,214 | 751 |
| 575–599 | 1,215 – 2,599 | 1,385 |
| 550–574 | 2,600 – 4,790 | 2,191 |
| 525–549 | 4,791 – 7,992 | 3,202 |
| 500–524 | 7,993 – 12,008 | 4,016 |
| 475–499 | 12,009 – 16,990 | 4,982 |
| 450–474 | 16,991 – 22,855 | 5,865 |
| 425–449 | 22,856 – 29,498 | 6,643 |
| 400–424 | 29,499 – 37,173 | 7,675 |
| 375–399 | 37,174 – 45,535 | 8,362 |
| 350–374 | 45,536 – 55,043 | 9,508 |
| 325–349 | 55,044 – 65,821 | 10,778 |
| 300–324 | 65,822 – 77,922 | 12,101 |
| 275–299 | 77,923 – 91,496 | 13,574 |
| 250–274 | 91,497 – 1,06,570 | 15,074 |
| 225–249 | 1,06,571 – 1,22,489 | 15,919 |
| 200–224 | 1,22,490 – 1,38,149 | 15,660 |
| 175–199 | 1,38,150 – 1,52,369 | 14,220 |
| 150–174 | 1,52,370 – 1,63,989 | 11,620 |
| 125–149 | 1,63,990 – 1,72,426 | 8,437 |
| 100–124 | 1,72,427 – 1,77,770 | 5,344 |
| 75–99 | 1,77,771 – 1,80,688 | 2,918 |
| 50–74 | 1,80,689 – 1,81,929 | 1,241 |
| 25–49 | 1,81,930 – 1,82,269 | 340 |
| 0–24 | 1,82,270 – 1,82,352 | 83 |
| -25–-1 | 1,82,353 – 1,82,354 | 2 |
| -50–-26 | 1,82,355 – 1,82,356 | 2 |
1,82,356 candidates appeared in 2022.
| Marks | Rank range | Candidates |
|---|---|---|
| 700–724 | 1 – 8 | 8 |
| 675–699 | 9 – 82 | 74 |
| 650–674 | 83 – 388 | 306 |
| 625–649 | 389 – 1,238 | 850 |
| 600–624 | 1,239 – 2,776 | 1,538 |
| 575–599 | 2,777 – 5,178 | 2,402 |
| 550–574 | 5,179 – 8,469 | 3,291 |
| 525–549 | 8,470 – 12,624 | 4,155 |
| 500–524 | 12,625 – 17,649 | 5,025 |
| 475–499 | 17,650 – 23,403 | 5,754 |
| 450–474 | 23,404 – 30,047 | 6,644 |
| 425–449 | 30,048 – 37,359 | 7,312 |
| 400–424 | 37,360 – 45,436 | 8,077 |
| 375–399 | 45,437 – 54,207 | 8,771 |
| 350–374 | 54,208 – 63,519 | 9,312 |
| 325–349 | 63,520 – 73,638 | 10,119 |
| 300–324 | 73,639 – 84,393 | 10,755 |
| 275–299 | 84,394 – 95,517 | 11,124 |
| 250–274 | 95,518 – 1,06,877 | 11,360 |
| 225–249 | 1,06,878 – 1,18,359 | 11,482 |
| 200–224 | 1,18,360 – 1,29,226 | 10,867 |
| 175–199 | 1,29,227 – 1,39,458 | 10,232 |
| 150–174 | 1,39,459 – 1,48,236 | 8,778 |
| 125–149 | 1,48,237 – 1,55,324 | 7,088 |
| 100–124 | 1,55,325 – 1,60,676 | 5,352 |
| 75–99 | 1,60,677 – 1,63,930 | 3,254 |
| 50–74 | 1,63,931 – 1,65,524 | 1,594 |
| 25–49 | 1,65,525 – 1,66,094 | 570 |
| 0–24 | 1,66,095 – 1,66,234 | 140 |
| -25–-1 | 1,66,235 – 1,66,248 | 14 |
| -50–-26 | 1,66,249 – 1,66,249 | 1 |
1,66,249 candidates appeared in 2021.
| Marks | Rank range | Candidates |
|---|---|---|
| 675–699 | 1 – 1 | 1 |
| 650–674 | 2 – 9 | 8 |
| 625–649 | 10 – 43 | 34 |
| 600–624 | 44 – 212 | 169 |
| 575–599 | 213 – 696 | 484 |
| 550–574 | 697 – 1,809 | 1,113 |
| 525–549 | 1,810 – 3,461 | 1,652 |
| 500–524 | 3,462 – 5,906 | 2,445 |
| 475–499 | 5,907 – 9,206 | 3,300 |
| 450–474 | 9,207 – 13,605 | 4,399 |
| 425–449 | 13,606 – 18,801 | 5,196 |
| 400–424 | 18,802 – 24,855 | 6,054 |
| 375–399 | 24,856 – 31,455 | 6,600 |
| 350–374 | 31,456 – 39,113 | 7,658 |
| 325–349 | 39,114 – 47,284 | 8,171 |
| 300–324 | 47,285 – 56,470 | 9,186 |
| 275–299 | 56,471 – 66,394 | 9,924 |
| 250–274 | 66,395 – 77,815 | 11,421 |
| 225–249 | 77,816 – 89,935 | 12,120 |
| 200–224 | 89,936 – 1,03,455 | 13,520 |
| 175–199 | 1,03,456 – 1,16,840 | 13,385 |
| 150–174 | 1,16,841 – 1,30,442 | 13,602 |
| 125–149 | 1,30,443 – 1,41,874 | 11,432 |
| 100–124 | 1,41,875 – 1,51,083 | 9,209 |
| 75–99 | 1,51,084 – 1,57,013 | 5,930 |
| 50–74 | 1,57,014 – 1,59,865 | 2,852 |
| 25–49 | 1,59,866 – 1,60,724 | 859 |
| 0–24 | 1,60,725 – 1,60,871 | 147 |
| -25–-1 | 1,60,873 – 1,60,876 | 4 |
1,60,875 candidates appeared in 2020.
How the NEET-PG Rank Predictor Works
Rank is a position within a cohort, not an absolute measure of performance. So the predictor does not try to map your marks straight onto a rank number. It first estimates your percentile — the fraction of candidates you finish ahead of — and only then converts that percentile into a rank using the expected size of the 2026 cohort.
That extra step matters because the NEET-PG cohort has grown steadily, from about 1.61 lakh candidates in 2020 to 2.30 lakh in 2025 — roughly 7.8% a year. Rank 10,000 in 2020 and rank 10,000 in 2025 are simply not the same achievement. Averaging past ranks directly would bake that drift into every prediction; working in percentile space removes it.
Why the curve never bends the wrong way: the model uses isotonic regression, which mathematically enforces that a higher score can never produce a worse rank. That guarantee is built into the method itself, not patched on afterwards.
How Accurate Is It? (Tested on 2025)
The model was trained only on 2020–2023 data and then tested against the 2025 result, which it never saw during fitting. That is a genuine out-of-sample test, not a score against its own training data.
| Actual rank band | Average error |
|---|---|
| Top 1,000 | ± 165 ranks |
| Top 5,000 | ± 301 ranks |
| Top 10,000 | ± 1,588 ranks |
| Overall | ± 6,181 ranks |
Accuracy is deliberately strongest at the top of the merit list, which is where a few hundred ranks change which branch and college are realistic. Lower down, absolute error grows but stays small relative to the rank itself.
Two Things Most Rank Predictors Get Wrong
- The 2020 paper was a different exam. NEET-PG 2020 had 300 questions worth 1200 marks; from 2021 it has been 200 questions worth 800. A raw 2020 score of 600 is not comparable to a 600 today. This model rescales 2020 marks onto the current 800-mark axis before using them.
- 2024 has no published marks. NBEMS released percentile and rank for 2024 but never released raw scores. It therefore cannot inform a score-to-rank model and is excluded from training rather than filled in with guessed marks.
Frequently Asked Questions
How accurate is this NEET-PG rank predictor?▼
The model is trained on NBEMS result data from 2020-2023 and tested against the 2025 result, which it never saw during training. On that holdout its median error was about 6,821 ranks overall, and roughly 165 ranks within the top 1,000. Accuracy is highest at the top of the merit list, where rank matters most, and loosens further down the list.
What score do I need for a rank under 10,000 in NEET-PG?▼
Based on 2020-2023 data mapped onto the current 800-mark pattern, a rank near 10,000 has historically needed roughly 520-545 marks, depending on the year's difficulty. Use the tool with your own score to see the estimated range rather than relying on a single cut-off number, because the score needed shifts every year with paper difficulty and candidate count.
Why does the predicted rank change if the number of candidates changes?▼
Rank is a position within a cohort, not an absolute measure. The model first estimates your percentile — what fraction of candidates you finish ahead of — which stays stable across years. It then converts that percentile into a rank using the size of the cohort. A larger cohort means the same percentile maps to a numerically larger rank, so the tool projects the 2026 cohort from recent growth and lets that projection be overridden once official figures are published.
Is NEET-PG 2020 data comparable to today's exam?▼
Not directly. NEET-PG 2020 had 300 questions worth 1200 marks, while 2021 onward has 200 questions worth 800 marks. A raw 2020 score is therefore on a different scale entirely. The model rescales 2020 marks onto the current 800-mark axis before using them, which is why the tool shows both the original and converted score when it displays a 2020 comparison.
Why is 2024 data not used to predict rank from score?▼
NBEMS published percentile and rank for NEET-PG 2024 but did not release raw marks. Without raw scores, 2024 cannot inform a score-to-rank relationship, so it is deliberately excluded from training and kept only as separate percentile-to-rank reference data. Using it as a score source would mean inventing marks that were never published.
Can a higher score ever give a worse rank in this tool?▼
No. The model uses isotonic regression, which mathematically enforces that rank never worsens as score increases. This is a structural property of the method rather than something checked after the fact, and it is covered by automated tests at both 1-mark and quarter-mark steps across the entire score range.
Does this predictor account for category, state quota or counselling?▼
No. It predicts your All India Rank in the common merit list only. Category rank, state quota rank and actual seat allotment depend on reservation, domicile, choice-filling and counselling round, none of which are part of NBEMS's published score-and-rank data. Treat the AIR here as the starting input to counselling research, not the final answer on seat prospects.
How much does the NEET-PG rank predictor cost?▼
It is a one-time payment of ₹49 that unlocks the tool for a full year. During that year you can re-check your rank as many times as you like, so you can re-run it after every mock test as your score improves. Access is restored by entering the same email you purchased with — there is no password to remember.