Checked 1 October 2026. The score scheme is from the CSE 2026 notice and 2026 question booklets. The CSE 2027 notification is not published yet; verify that it retains the scheme when issued. The calculator gives raw arithmetic only.
UPSC Civil Services Preliminary Examination has two compulsory objective papers. General Studies Paper-I has 100 questions for 200 marks; General Studies Paper-II, known as CSAT, has 80 questions for 200 marks. Each is two hours. CSAT is qualifying at 33%, which is 66 marks out of 200 by our arithmetic. For both papers, a wrong answer loses one third of the marks assigned to that question; a blank loses nothing. A candidate must appear in both papers. These are the current 2026 rules to prepare against until the 2027 notice appears.
UPSC CSE 2026 notice and actual 2026 paper booklets; 66 is our 33% arithmetic.
The paper roles are different
| Paper | Questions | Maximum | Per correct | Per wrong | Selection role |
|---|---|---|---|---|---|
| General Studies I | 100 | 200 | 2 | −2/3 ≈ −0.67 | Category cut-off for Main shortlist |
| General Studies II (CSAT) | 80 | 200 | 2.5 | −5/6 ≈ −0.83 | Qualifying at 33% = 66 marks |
Prelims raw-score checker
Enter correct and wrong answers. Blank items have no penalty. This shows raw arithmetic, not normalisation or an official result.
GS-I: +2 correct, −2/3 wrong. CSAT: +2.5 correct, −5/6 wrong. CSAT needs 33% of 200 = 66. Only UPSC decides the year’s GS-I cut-off.
The calculator uses fractions before rounding the display. That matters near a threshold: the exact one-third penalty is two-thirds of a GS-I question’s two marks, and five-sixths of a CSAT question’s 2.5 marks. If you subtract a rounded 0.67 or 0.83 after every wrong answer, repeated rounding can create a small discrepancy. UPSC is the authority for official marks; the tool is a practice aid using the printed marking rule. It does not account for cancelled questions, final answer keys or any later change of rule.
GS-I determines the competitive Preliminary cut-off after CSAT qualification. The 2025 General GS-I threshold was 92.66, but that is a historical outcome, not the 2027 target. The 2026 cut-off had not been published by 1 October 2026 because UPSC says it will release marks, cut-offs and answer keys after the whole CSE 2026 process ends. A coaching-site “2026 safe score” is therefore an estimate, however plausible it looks. Keep that label when comparing your practice marks.
CSAT is not added to GS-I to create a 400-mark merit score. A candidate with 120 in GS-I and 64 in CSAT fails the published 33% CSAT condition; the high GS-I figure cannot compensate. Conversely, a candidate with 110 in CSAT and a GS-I score below the category threshold cannot use CSAT surplus to cross that competitive gate. This two-gate structure is why a Prelims study plan must measure the two papers separately.
Worked raw-score examples
| Paper | Correct | Wrong | Blank | Raw arithmetic |
|---|---|---|---|---|
| GS-I example A | 50 | 20 | 30 | 50×2 − 20×(2/3) = 86.67 |
| GS-I example B | 55 | 15 | 30 | 55×2 − 15×(2/3) = 100.00 |
| CSAT example A | 35 | 15 | 30 | 35×2.5 − 15×(5/6) = 75.00 |
| CSAT example B | 30 | 15 | 35 | 30×2.5 − 15×(5/6) = 62.50 |
The examples are our arithmetic illustrations, not actual UPSC candidate scores. In GS-I example A, 50 correct answers yield 100 marks and 20 wrong answers deduct 13.33 recurring, leaving about 86.67. A difference of five additional correct answers and five fewer wrong answers in example B raises the raw total to 100. That is not merely a question-count gain: fewer penalties also help. In a real paper, the final answer key determines which responses are counted as correct, and questions may be treated differently after objections.
In CSAT example A, 35 correct and 15 wrong produce 75, above the 66 qualifying line. Example B, with 30 correct and 15 wrong, yields 62.5, below it. Both examples leave many questions blank. They show that “attempt 50 questions” alone is not a useful pass rule; accuracy changes the result. They also show why a candidate sitting near 66 should not rest on an approximate coaching answer key. Build a margin in timed practice and wait for the official outcome rather than treating a half-mark estimate as certain.
What one-third negative marking means for guessing
| Choices after elimination | Chance correct if equally likely | Expected gain, question-mark units |
|---|---|---|
| No option eliminated from four | 1/4 | 0 under the one-third penalty |
| One option eliminated | 1/3 | 1/9 |
| Two options eliminated | 1/2 | 1/3 |
The 2026 notice says each objective question has four alternatives. If a candidate has no information and assumes each option is equally likely, the expected value of a blind guess under a one-third penalty is zero: one-quarter chance of earning one question’s marks and three-quarters chance of losing one third of them. This is our probability calculation, not a promise about a particular paper. If the candidate can genuinely eliminate an option, the mathematical expected value becomes positive under that simplified model. But an imagined elimination is not information; forcing a choice from an uncertain hunch can still lower an actual score.
The right practical use of this arithmetic is to diagnose answer-selection habits in mocks. Record questions you knew, questions where you eliminated options for a stated reason, and questions marked almost at random. After checking the official or credible answer key, see whether the second category really adds marks for you. Avoid a universal command such as “attempt exactly 85 GS questions.” The efficient attempt count depends on paper difficulty, your knowledge and time. The penalty makes accuracy and selection inseparable.
Multiple marked options count as a wrong answer, according to the notice. An unmarked question has no penalty. UPSC uses an OMR sheet for Prelims and requires a black ball-point pen in the 2026 instructions. Those mechanical details matter because an otherwise correct thought not properly marked cannot be recovered by a blog calculator. Read the actual admit-card and booklet instructions every cycle; do not assume an online CBT interface for this exam.
CSAT preparation when it is “only qualifying”
Qualifying does not mean optional. The paper has its own two-hour sitting and passing line. Candidates comfortable with one CSAT topic can still be exposed by another: timed comprehension, reasoning and numerical problem-solving all reward practice. Begin with an actual past UPSC CSAT paper to find which question types consume time and which errors are conceptual. Then practise to a stable margin above 66 rather than trying to land exactly on it. A narrow mock pass can disappear through answer-key differences or a few avoidable OMR errors.
Set a CSAT maintenance routine throughout GS-I preparation. One timed session and a separate review of wrong answers can be more valuable than reading another generic “CSAT tips” list. If your mock raw score is consistently below 66, assign more time until the failure mode changes. If it is consistently well above, keep periodic full-paper checks so a future gap does not go unnoticed. This is our study advice; UPSC prescribes the qualifying rule, not a study frequency.
As 2027 approaches, compare the notification’s Appendix on Preliminary rules with this guide. The CSE 2027 notice could alter details. Update the calculator only after reading an official change. The current 2026 document and the 2026 question booklets support the figures here, but a 2027 candidate should not rely on a year-old website alone for the final exam-day instructions.
A simple diagnostic before exam day
Run the calculator on three timed papers rather than one. Write down the GS-I and CSAT raw marks, the number of blank questions, and the reason for every error. If a paper would qualify in CSAT only by one or two marks, do not call that a secure pattern: another set of questions or a different answer key can move it. If the same subject repeatedly causes wrong answers, repair that subject before raising your attempt count. If a long calculation repeatedly costs time, rehearse a quicker route or leave that item for a second pass. The purpose of this small audit is to choose a preparation action from evidence, not to predict a UPSC result. The official answer key and category cut-off remain decisive after the examination cycle.
What a good mock review should record
Track correct, wrong and blank counts separately for GS-I and CSAT, then the raw score by the exact penalty. Add a column for why the wrong answer happened: unfamiliar fact, misread stem, calculation, time pressure or speculative guess. The aggregate mark tells you whether you are near a threshold; the error labels tell you what to fix. A chart of only score totals can hide the fact that two identical marks came from very different attempt strategies.
If a GS-I mock score looks strong but the CSAT score is weak, do not combine them to reassure yourself. If CSAT is comfortable but GS-I is below several past category thresholds, focus on GS-I coverage and question selection. Our companion cut-off guide prints every official 2019–2025 category row so you can place raw GS-I practice in historical context, while remembering a mock is not an official UPSC paper. A third-party answer key should always be labelled provisional until UPSC publishes its key after the cycle.
Finally, keep the attempt rule in mind. The CSE notification counts a candidate as having used an attempt if they actually appear in even one Preliminary paper. Merely applying without appearing does not consume an attempt. That is a serious decision for candidates close to a category attempt limit. It does not change the obligation to sit both papers once you choose to compete: an appearance in only one makes you ineligible for Main while still consuming an attempt under the published rule.
What are the CSAT qualifying marks?
33% of 200, which is 66 marks by arithmetic under the current UPSC scheme.
How many questions are in CSAT?
The 2026 question booklet has 80 questions for 200 marks.
How many questions are in GS-I?
The 2026 GS Paper-I booklet has 100 questions for 200 marks.
What is the negative marking?
One third of the marks assigned to a wrong question: two-thirds of a mark-equivalent GS-I item worth 2, and five-sixths for a CSAT item worth 2.5.
Are CSAT marks added to GS-I for the cut-off?
No. CSAT is qualifying at 33%; the competitive Preliminary cut-off is based on GS-I.
Does a blank question lose marks?
No. UPSC applies no penalty when no answer is marked.
Does appearing in one Prelims paper use an attempt?
Yes. The CSE notice treats actual appearance in even one Preliminary paper as an attempt, though both papers are required for qualification.
Sources
- official — UPSC — Civil Services Examination 2026 notice (revised, 6 February 2026)
- official — UPSC — CSE 2025 cut-off marks (9 March 2026)
- official — UPSC — Civil Services Preliminary Examination 2026 result (15 June 2026)
official = a document published by the conducting body. reported = a news or coaching site we could not check against an original. Where sources disagree this page says so rather than picking one.
