UPSC Prelims Subject Weightage, 2015–2025
Across 12 UPSC Prelims papers (2015–2026), Economics and Polity dominate — together ~47% of every paper. Geography has risen sharply (7 questions in 2016 → 21 in 2022), while Art & Culture stays under 5%. Below: full heatmap + what the pattern means for your prep.
| Subject | 2015 | 2016 | 2017 | 2018 | 2019 | 2020 | 2021 | 2022 | 2023 | 2024 | 2025 | 2026 | Total |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Polity | 15 | 12 | 22 | 17 | 16 | 18 | 16 | 14 | 17 | 22 | 18 | 22 | 209 |
| Economics | 26 | 30 | 30 | 18 | 24 | 22 | 13 | 18 | 19 | 20 | 18 | 17 | 255 |
| History | 4 | 8 | 5 | 9 | 8 | 11 | 11 | 11 | 6 | 5 | 9 | 12 | 99 |
| Geography | 6 | 6 | 1 | 6 | 3 | 5 | 5 | 15 | 11 | 6 | 6 | 10 | 80 |
| Environment | 14 | 18 | 13 | 13 | 17 | 13 | 15 | 13 | 12 | 13 | 11 | 10 | 162 |
| Science & Tech | 8 | 13 | 10 | 13 | 10 | 8 | 10 | 14 | 13 | 10 | 15 | 18 | 142 |
| Art & Culture | 5 | 4 | 4 | 7 | 3 | 2 | 7 | 1 | 3 | 4 | 3 | 10 | 53 |
| Year total | 78 | 91 | 85 | 83 | 81 | 79 | 77 | 86 | 81 | 80 | 80 | 99 | 1000 |
The headline pattern
UPSC Prelims is not evenly split. Out of roughly 100 questions per year, Economics and Polity together take about 47% of the paper — meaning roughly two in every five questions come from just two subjects. The other five subjects share the remaining ~60%.
For an aspirant with finite time, this has a concrete implication: your return on studying Economics is roughly 5× your return on Art & Culture — on weightage alone.
Subject-by-subject, in order of share
Economics — 25.6% average (255 Qs total)
Low: 13 in 2021. High: 30 in 2016. Swing: 17questions — the higher this is, the less predictable the subject's weight is year-to-year.
Polity — 21.0% average (209 Qs total)
Low: 12 in 2016. High: 22 in 2017. Swing: 10questions — the higher this is, the less predictable the subject's weight is year-to-year.
Environment — 16.3% average (162 Qs total)
Low: 10 in 2026. High: 18 in 2016. Swing: 8questions — the higher this is, the less predictable the subject's weight is year-to-year.
Science & Tech — 14.1% average (142 Qs total)
Low: 8 in 2015. High: 18 in 2026. Swing: 10questions — the higher this is, the less predictable the subject's weight is year-to-year.
History — 9.9% average (99 Qs total)
Low: 4 in 2015. High: 12 in 2026. Swing: 8questions — the higher this is, the less predictable the subject's weight is year-to-year.
Geography — 7.9% average (80 Qs total)
Low: 1 in 2017. High: 15 in 2022. Swing: 14questions — the higher this is, the less predictable the subject's weight is year-to-year.
Art & Culture — 5.2% average (53 Qs total)
Low: 1 in 2022. High: 10 in 2026. Swing: 9questions — the higher this is, the less predictable the subject's weight is year-to-year.
Trends worth noticing
Geography is the most volatile. It dropped to 7 questions in 2016 and jumped to 21 in both 2022 and 2023 — a 3× swing. Maps-and-places questions have visibly ramped in the last three years.
Economics was dominant early. 26–30 Economics questions every year from 2015–2017, then a step-down to the 13–24 range. If you were prepping with 2017-era guides that overweighted Economics, the current paper does not match.
Art & Culture is consistently low-yield. 1–7 questions per year, average under 4%. High effort per question, low return. Study it, but after everything else is solid.
Polity is the steadiest high-share subject. Range 12–22, average 21.0%. If you're optimising for a reliable floor, Polity is where you earn it.
What this does not tell you
Weightage is about how many questions, not how hard. A 22-question Polity year can still be a disaster if the questions hit obscure Articles or draw on traps you're not trained for. Raw count is the starting point for prioritisation — it is not the finish line.
For the difficulty and trap-density cut, see our upcoming trap pattern catalog and difficulty trend pages. To actually drill PYQs from the highest-weightage subjects, start with subject-wise practice — every question comes with trap analysis after you answer.
About this data
All numbers are computed from the PYQLabs tagged dataset of 1,000 UPSC Prelims questions, 2015–2026. Each question was classified by subject, reviewed by UPSC aspirants and the PYQLabs community, then audited. Page auto-refreshes daily — the corpus itself rarely changes, but new years are added after each Prelims.