What Percentage of UPSC Aspirants Actually Clear It?

Someone posted their birth details in a Careers360 thread once and asked, basically, "will I clear IAS." The answer they got back was blunt: success ratio is around 0.3 percent, millions apply every year, now ask yourself if you can suffer that level of competition. Good luck.

That's the tone most people absorb about UPSC's odds: a number designed to intimidate, delivered like a dare. I wanted the actual number, not the dare. So I went and checked it against what UPSC itself has published, not what a coaching Instagram post says the number is.

The straight answer

Since 1947, roughly 60.4 million people have applied across various UPSC exams, and the cumulative success rate over that entire period works out to about 0.53%. The annual rate has fallen sharply over time, from around 5% in 1947 to roughly 0.2% by 2022, driven mostly by the explosion in applicant numbers, not by the exam getting proportionally harder. For CSE 2025 specifically, roughly 937,876 people applied, down from a peak near 1 million the year before.

Where the "0.2%" number actually comes from, and why it's misleading

The 0.2% figure isn't fabricated. It's a real, sourced number from UPSC's own historical trend. But it's doing something sneaky by being reported alone, without the applicant-count context that explains it. The annual success rate has dropped mainly because far more people are applying now than in 1947, not because the exam has become proportionally harder to clear, and not because fewer people are getting selected in absolute terms. Vacancies have stayed roughly in the same broad range for years; applicants have multiplied many times over. The ratio moved because the denominator grew, not because the numerator shrank.

That distinction matters enormously for how you should feel reading the number. "0.2% clear it" sounds like an act of God. "Vacancies stayed roughly stable while ten times as many people applied" sounds like a crowded room, which is a genuinely different and more actionable thing to think about.

The real funnel, stage by stage

Here's where the number actually gets filtered, and it's a lot less mysterious once you see each stage:

Stage	What happens	Rough proportion
Total applicants	Everyone who submits a form	100%
Actually appear for Prelims	Roughly half show up at all	~50-55%
Clear Prelims to Mains	Objective-type screening round	~10-12% of those who appear
Clear Mains to Interview	Written stage across 9 papers	~20-25% of Mains-qualified
Final selection	After Personality Test	800-1,200 candidates in a typical year

The single biggest and most under-discussed filter in this whole chain is the second row: only about half of the people who apply actually show up to sit the Prelims exam. For CSE cycles with around 13 lakh applicants, that means only 6 to 7 lakh people even walk into a test centre. That "absentee phenomenon" alone roughly doubles your realistic odds compared to the raw applicant-to-vacancy ratio most headlines quote, because half the stated competition was never real competition to begin with. It's people who applied out of curiosity, family pressure, or as a backup, and never seriously intended to appear.

What the data says about who actually wins

UPSC's own Annual Report data, once you dig into it rather than just the headline ratio, tells a genuinely counterintuitive story about who succeeds:

Only about 6.2% of selected candidates clear it on their very first attempt. The single highest-success attempt number is the fourth: roughly 22% of selections come from candidates on their 4th try, not their 1st. The best-represented age band among selected candidates is 24 to 26. And about 71% of those selected come from a science educational background, not commerce or arts, despite a large share of aspirants coming from commerce and humanities streams.

None of that means a first attempt is doomed or a commerce background is a disadvantage severe enough to abandon the exam over. Plenty of people do succeed on attempt one, and plenty from non-science backgrounds get selected every year. What it does mean is that the exam rewards accumulated preparation and exam-specific temperament more than raw academic background, and that a failed first or second attempt is not the signal of doom it's often treated as inside a batch group where everyone's silently comparing attempt numbers.

Think about what the 4th-attempt spike is actually describing. It's not that trying four times magically works. It's that by the fourth attempt, a serious candidate has typically been through three full Prelims-Mains cycles, which means three rounds of seeing where they actually lost marks, three rounds of adjusting an optional subject choice or an answer-writing style that wasn't working, and three interview cycles' worth of feedback if they got that far even once. That's not luck compounding. It's a genuinely different, more refined candidate showing up to attempt four than showed up to attempt one. The exam's format, with its heavy weight on subjective Mains answers and a Personality Test that rewards a specific kind of composed articulation, seems to reward exactly this kind of iterative refinement more than it rewards raw first-attempt talent.

The science-background number deserves a similarly careful read. It doesn't mean the syllabus favours science students on content. General Studies and the optional subjects are deliberately built to be accessible across educational backgrounds. What it more likely reflects is a selection effect upstream of the exam itself: the kind of student who pursues UPSC seriously after a science degree may already be self-selecting for a certain risk tolerance, discipline in structured preparation, or family expectation pattern that correlates with eventual success, not that a physics or engineering degree itself teaches you to write a better Mains answer on governance or ethics.

Where people get this decision wrong

The most common mistake is treating the headline 0.2% as a personal probability rather than a population-level statistic that includes tens of thousands of non-serious or unprepared applicants who were never really in your competitive set. If you're someone who's actually studying with structure, showing up for mocks, and tracking your own Prelims performance against real cutoffs, your functional competition isn't the full applicant pool. It's the smaller set of people who are similarly serious. That's a meaningfully different, though still genuinely tough, number to be reasoning about.

The second mistake is the opposite error: using "the odds get better with each attempt" data to justify grinding out attempt after attempt without changing anything about the approach. The 4th-attempt success spike reflects candidates who used their earlier attempts to genuinely improve (refine their optional subject choice, fix a weak Mains paper, get calmer in interview), not candidates who simply repeated the same preparation four times and eventually got lucky. If your Prelims score hasn't moved between attempts, more attempts alone won't fix that; something in the approach needs to.

The third mistake is comparing your own single data point against an average that was never built to predict any one person's outcome. Averages describe a population. They don't know your specific optional subject, your specific writing practice, or how many actual hours you're putting in versus how many you're logging in a study-tracker app for the reassurance of the number going up.

There's a fourth mistake worth naming too, because it shows up constantly in how this data actually gets used in batch groups and comparison threads: treating a friend's or senior's attempt number as directly comparable to your own without accounting for category differences. A General category candidate on their 5th attempt is in a genuinely different, more urgent position than an SC/ST candidate on their 5th, simply because one has a hard 6-attempt ceiling closing in and the other doesn't. The 22%-on-4th-attempt statistic is pooled across all categories and doesn't tell you anything about whether your specific 4th attempt, given your specific remaining runway, should look and feel like anyone else's.

The actual reframe

The honest version of this number isn't "you have a 0.2% chance" or "you have a great chance, don't worry." It's this: out of everyone who applies, roughly half aren't real competition because they never show up. Of the ones who do show up and treat this seriously, the funnel is still genuinely brutal. But it's a funnel that has consistently rewarded a specific kind of compounding, patient preparation over raw first-attempt brilliance, going by UPSC's own selected-candidate data.

That's a harder story to fit on an intimidating Instagram graphic. It's also the one that's actually true.

Where Artha fits

The scariest part of this number isn't the math. It's sitting with it alone at 2am after a mock test went badly, doing the mental arithmetic on how many real attempts you might have left. UPSC Talk on Artha exists for exactly that conversation, with someone else who's also counting, not performing confidence for a batch group.

Closing

The number was never really 0.2% for you specifically. It was 0.2% for everyone, including people who were never seriously competing. So which number are you actually up against, once you take out everyone who was never really in the race?