UPSC CSE Cut-Off 2019–2025: Prelims, Mains, Final by Category

UPSC’s Preliminary threshold is based on GS Paper-I. GS Paper-II, commonly called CSAT, has a separate qualifying condition of 33% and does not add to the GS-I cut-off. A candidate with high GS-I marks who falls below the CSAT minimum cannot use the GS-I lead to compensate. Likewise, the “Prelims cut-off” is not out of 400 just because the two papers total 400 available marks. That is perhaps the most common source of bad comparisons in search snippets and coaching graphics.

The 2023 General line was 75.41; 2024 rose to 87.98 and 2025 to 92.66. The spread from 2023 to 2025 is 17.25 marks, our subtraction from UPSC’s published values. It does not establish that the exam became easier or that a given study method worked better. Cut-offs reflect the paper, cohort, vacancies and selection rule. A candidate should use the series to understand uncertainty and to test whether a mock score would have been competitive in several historical fields, not to declare that a future mark is guaranteed.

Category labels matter too. OBC in the UPSC table refers to the eligible reservation category under the applicable notice, not every person with an OBC certificate of any form. PwBD is split into subcategories, which the official PDF labels PwBD-1, PwBD-2, PwBD-3 and PwBD-5. Merging them into one “PwD cut-off” discards real variation. For example, in 2025 the published PwBD-1 Preliminary figure was 76.66 while PwBD-2 was 54.66. If you are checking your own category, read the exact classification in the notice and the PDF rather than picking the lowest disability line from an unsourced table.

The written Main threshold is reported out of 1,750 merit marks. The qualifying Indian-language and English papers are separate conditions; the interview is not included yet. In the 2025 cycle, the General Main figure was 739. That is not a target per paper, and dividing by the number of papers would create a misleading “minimum average.” UPSC sets qualifying and merit rules at the paper level as well as an aggregate shortlist. The published cut-off is the last qualifying score at that stage under the official category classification.

The General written Main series is comparatively compact—751 in 2019, 736 in 2020, 745 in 2021, 748 in 2022, 741 in 2023, 729 in 2024 and 739 in 2025. It still spans 22 marks. Two candidates with the same aggregate may differ on qualifying-paper status or other rules. A general candidate who sees 739 should therefore not assume that writing exactly 739 in a later cycle earns an interview. The field changes and UPSC decides the cut-off after evaluation. Even for the same year, a score at the threshold describes the last selected category-specific candidate, not a universal pass line independent of the process.

Use the written Main table to interpret historical performance if you have an official mark sheet. Compare the same cycle, same stage and same category. Do not compare a 2025 mock scored by a coaching evaluator with UPSC’s 2025 Main figure as if both were calibrated assessments. A mock may still help identify answer-writing weaknesses, but its number is not in UPSC’s distribution. When you discuss a Main score with peers, state whether it is a self-estimate, evaluated mock or official marks released by the Commission.

The final threshold uses the 1,750 written merit marks plus the 275-mark Personality Test. The 2025 General final cut-off was 963. That is a published result of the completed 2025 cycle. It is not the score that UPSC required to be called to interview; that earlier gate was the written Main line of 739. A chart that lists 963 as a “Mains cut-off” confuses the two steps and can make candidates believe they missed an interview by more than they actually did.

One way to see the role of the interview is to compare the Main and final scales. If a hypothetical candidate earned 739 in written Main and 224 in the Personality Test, the sum would be 963. That is our arithmetic illustration, not an assertion about the last recommended candidate’s actual marks or profile. The real final result depends on each individual’s combined marks and category, and the published cut-off is an outcome of the ranked list. UPSC’s final PDF does not permit us to infer every selected candidate’s interview score from the three threshold rows.

The General final series from 2019 through 2025 ranges from 944 to 963 in the official PDFs, a 19-mark span by our calculation. Other categories move differently, and PwBD subcategories can shift more sharply because their candidate pools and vacancy counts differ. Do not make a causal story from a single row. If you publish a comparison, show the full year and stage and link the actual UPSC PDF. That source discipline is more useful than a confident “expected cut-off” number that will be forgotten when the result arrives.

UPSC’s CSE 2026 Preliminary result note says marks, cut-off marks and answer keys will be uploaded after the entire examination process is over, that is, after the final result. The written Main result itself had not appeared in the official source set checked by 1 October. Therefore a page claiming to show the “official UPSC Prelims 2026 General cut-off” now must be checked very carefully: it may be a coaching estimate, a threshold from another year or an unsupported figure. Labelled estimates can be useful for planning, but they should not be placed into an official historical table.

If you are a 2026 candidate, the published Preliminary result tells you whether you advanced. If you are planning for 2027, the 2019–2025 table is enough to set a robust range for practice. Aim to build a clear margin above past thresholds while independently keeping CSAT safe. That is our strategic recommendation, not a 2027 predicted cut-off. The 2027 notification, paper and candidate field do not yet exist; no honest publisher can convert this historical dataset into certainty about them.

Ask five questions. Does it link to a UPSC PDF? Does the file name and heading identify the CSE cycle? Does the row say Preliminary, written Main or final? Does it give the marks denominator? Does it preserve EWS, OBC, SC, ST and individual PwBD columns? If any answer is missing, treat the graphic as an interpretation until verified. A citation to UPSC’s home page is weaker than a link to the specific cut-off PDF, because the exact number should be checkable in one click.

Also watch for mixing “minimum qualifying marks” with a test’s qualifying rule. CSAT’s 33% requirement is a fixed rule in the notification, while the GS-I category cut-off is a result-dependent threshold. Main’s qualifying language papers and paper-wise floors are different from the aggregate written Main cut-off. Each has a different purpose. The tables above keep the three published threshold types separate and are intended to be read alongside the selection process, not in place of it.