UPSC Prelims Cut-Off Trends: What the Numbers Teach You
Updated 14 August 2026
UPSC does not announce a pass mark for Prelims in advance. The cut-off for any year is simply the General Studies Paper-I score of the last candidate who qualified for Mains in each category — and the Commission publishes it only after the entire cycle, Interview included, is over. The 2024 cut-offs came out in April 2025; the 2025 figures arrived in March 2026. Line up seven years of this data, though, and the numbers teach some blunt lessons about how this exam actually behaves.
Official GS Paper-I cut-offs, 2019–2025
All figures are out of 200 and are computed on GS Paper-I alone. CSAT (GS Paper-II) is qualifying at 33%, and each wrong answer in GS-I costs one-third of the question's two marks.
| Year | General | EWS | OBC | SC | ST |
|---|---|---|---|---|---|
| 2019 | 98.00 | 90.00 | 95.34 | 82.00 | 77.34 |
| 2020 | 92.51 | 77.55 | 89.12 | 74.84 | 68.71 |
| 2021 | 87.54 | 80.14 | 84.85 | 75.41 | 70.71 |
| 2022 | 88.22 | 82.83 | 87.54 | 74.08 | 69.35 |
| 2023 | 75.41 | 68.02 | 74.75 | 59.25 | 47.82 |
| 2024 | 87.98 | 85.92 | 87.28 | 79.03 | 74.23 |
| 2025 | 92.66 | 89.34 | 92.00 | 84.00 | 82.66 |
Figures as disclosed by UPSC after each year's final result. For the official record, always verify at upsc.gov.in.
What the trend actually says
- The cut-off tracks paper difficulty, not candidate quality. General moved from 98.00 in 2019 down to 75.41 in 2023 and back up to 92.66 in 2025. The applicant pool did not change dramatically in those years — the papers did. A "low cut-off year" is not an easy year; it is a hard paper that pulled everyone's score down together.
- The working band is roughly 75–98. In seven years the General cut-off never left this range. Any preparation plan built around one specific number inside that band is a gamble.
- The OBC and EWS gaps have nearly closed. In 2024, OBC stood at 87.28 against General's 87.98 — a gap of 0.70 marks. In 2025 it was 0.66. At the Prelims stage, reservation offers almost no cushion; every serious aspirant should prepare against the General benchmark.
- SC and ST cut-offs swing hardest. ST fell to 47.82 in the brutal 2023 paper, then climbed to 82.66 by 2025 — a 35-mark swing in two years. Past-year cut-offs are the least reliable target for exactly the categories that lean on them most.
Lesson 1: Prepare for the ceiling, not the average
On exam day you cannot know whether you are sitting a 75-cut-off paper or a 98-cut-off paper. The only rational response is to target the top of the band plus a margin — a consistent 105–110 in serious full-length mocks. That covers the hardest year on record and absorbs the 10–15 marks of exam-day slippage that nerves, ambiguous questions, and time pressure extract from almost everyone.
Lesson 2: Accuracy beats attempt count
The negative-marking arithmetic is unforgiving. Take two candidates who each get 60 questions right (120 marks). One attempted 90, so 30 wrong answers cost 20 marks: net 100 — below the General cut-off in five of the last seven years. The other attempted 80, losing only about 13 marks: net roughly 106.7 — above the cut-off in five of those years. Same knowledge, nearly seven marks apart, purely on the decision to skip ten blind guesses. The skill to drill in mocks is not attempting more; it is recognising which questions to leave. A sound working rule: attempt when you can confidently eliminate two options, leave pure coin-flips alone.
Lesson 3: Treat CSAT as a real paper
The table above has a silent filter: your GS-I score is counted only if you cross the 33% qualifying line in CSAT — roughly 66 marks out of 200. Since 2021, CSAT papers have become noticeably tougher, leaning harder on quantitative aptitude and multi-step reasoning, and every cycle sees aspirants with comfortable GS-I scores eliminated by the qualifying paper. If maths is not your strength, one dedicated CSAT practice session a week from the start of preparation is cheap insurance. Discovering the problem in the final month is how attempts get wasted.
Using this data in your mock-test routine
- Benchmark against 105+, not last year's cut-off. Chasing 88 because 2024 closed at 87.98 leaves you exposed to a 2019-style paper.
- Track accuracy as its own metric. Score can rise while accuracy falls — a fragile improvement. Aim for 75–80% accuracy on attempted questions before pushing attempt counts higher.
- Write CSAT mocks in the same sitting as GS-I, in the afternoon slot, so fatigue is part of the simulation.
- Ignore "expected cut-off" chatter after the exam. The official figure arrives only with the final result, months later. Candidates who spent that window preparing for Mains — instead of refreshing prediction threads — are the ones the next table quietly records.
