UPSC Mains GS2 2025 Analysis: PYQ Patterns Decoded

The UPSC Mains GS Paper 2 2025 turned out to be, in Neil Sir's words, almost a remix of previous year questions, so much so that it felt like the 2019 paper reworded for 2025. In this live breakdown he walks through all 20 questions of the GS2 2025 paper and shows how previous year question (PYQ) analysis, basic standard sources and plain common sense are enough to generate points on each one. The takeaway for Mains aspirants is simple: master the basics, recognise the repeated themes, and back them with answer writing practice.

Key takeaways

• The 2025 GS2 paper was largely a reflection of earlier papers, with several questions (RPA, the Attorney General, the India-Africa digital partnership) being near copy-paste of 2019.
• Most polity and governance questions are fully covered by basic sources like Laxmikant; the difficulty usually lies in the complicated wording, not the content.
• When you lack deep, specific knowledge, speak in terms of facts and a balanced both-sides analysis rather than asserting an unsupported opinion.
• A few thought-provoking questions, such as e-governance's technology bias and energy security in foreign policy, are the real differentiators between an average aspirant and one who scores exceptionally well.
• For most questions the safe conclusion is a balanced one: two institutions must work "in tandem", a power is "strong but not supreme", and so on.
• For 2026 aspirants the action point is to brainstorm each of these 20 questions, write them out, and build the habit of generating points from basics plus common sense.

Why GS2 2025 felt like a remix of previous papers

Neil Sir's central thesis is that PYQs hold the key to the exam, and GS2 2025 proves it. Question after question echoes themes that have appeared before. The Attorney General question (Article 76, advising and appearing for the government, participating in Parliament with no voting power) is described as a copy-paste of 2019. The India-Africa digital partnership question is again traced back to 2019. Constitutional morality, the amending power, the collegium, federalism and the national commissions are all themes that have been asked repeatedly.

His point is that sometimes only the wording changes; the wording is what makes a basic question seem more intellectual under exam pressure. Once you comprehend a question correctly, the "fogginess" disappears and you realise it is asking for content you already know.

How to generate points when you lack deep knowledge

The method Neil Sir demonstrates is to lean on three things: PYQ familiarity, basic sources, and common sense, finished with answer writing practice.

Speak in facts, not opinion

On the RPA question about whether a disproportionate increase in a legislator's assets amounts to undue influence and therefore a corrupt practice, he warns against being too assertive just because the question is framed that way. Detailed knowledge here is hard, so speak in terms of facts: disproportionate assets are bad, but by themselves they are not a corrupt practice, because they are already covered by the Prevention of Corruption Act, the Benami Act and the PMLA. The RPA needs that support.

Aim for a balanced conclusion

Many questions resolve to a both-sides analysis with a measured ending. Administrative tribunals exist for expertise and speed in technical disputes, but tribunals and courts must work in tandem. The amending power is strong but not supreme, and the core of the Constitution remains inviolable. Constitutional morality, anchored in liberty, equality, fraternity and dignity (seen in the Navtej Singh Johar and Sabarimala cases), is the guardrail that lets judicial independence and accountability go hand in hand.

Polity and governance: the core that comes from the basics

A large block of the paper is straightforward polity, all of it traceable to standard sources:

• President's pardoning power compared with the US: in India the President acts on the advice of the Council of Ministers, with commutation and remission subject to judicial review, while the US President acts independently and can grant a pre-emptive pardon even before conviction.
• Jammu and Kashmir as a Union Territory with a legislature after 2019, where the Lieutenant Governor holds significant power, and the friction seen in Delhi between the LG and the Council must be avoided.
• The collegium system's evolution (first and second judges cases, the striking down of the NJAC), contrasted with US appointments where the President proposes and the Senate confirms; India gains insulation from politics but suffers opacity, the US gains transparency but invites politicisation.
• Centre-state financial relations and fiscal federalism, where the 14th Finance Commission gave a 42% share, GST created friction (with petrol and alcohol kept outside it), and the 15th Finance Commission and disaster funds brought further change. The conclusion is cooperation, with both sides coming to the same table.

The differentiator questions

Two questions stand out as the ones that separate strong candidates.

The e-governance question argues these projects have a built-in bias towards technology and back-end integration rather than user-centric design. Neil Sir calls this a genuinely thought-provoking, analysis-based question. Build the answer with examples like repeated Aadhaar OTP failures, language accessibility gaps, KYC issues, poor grievance visibility, weak cyber security and untrained field staff, while conceding that things are improving. The line to land: build systems around people, not servers; technology should be an enabler, not the end game.

The energy security question is described as the first forward-looking, GS4-style question in the paper. Its prelude mentions the Middle East, but the question itself is not Middle East specific, so the answer must be holistic. India imports roughly 85% of its oil and around 45 to 50% of its gas, so the response should cover long-term contracts, diversification (Russia, the US, Latin America), buying equity for ownership, securing sea lanes under the "security and growth for all in the region" idea, building refining hubs, leveraging technology and diaspora, and treating it all as a give-and-take in foreign policy.

Social justice and international relations themes

The social justice and IR questions again reward connecting basics to the syllabus. Women's social capital (networks and trust, illustrated by SHGs, Kudumbashree and Amul) advances empowerment if it is democratised across society rather than captured by one elite group. CSOs are by and large a complement to the state, not anti-state actors, provided FCRA and foreign-funding rules stay transparent. The NCPCR question on children in the digital era points to online grooming, cyber bullying, privacy breaches, addictive gaming, deep fakes and the digital divide, against existing frameworks like POCSO, the Juvenile Justice Act and the IT Rules. The closing UN reform question turns on the East-West imbalance, the P5 veto crisis, G4 aspirations and the Global South's climate demands, with the warning that without reform the UN risks irrelevance.

Who should watch this

This breakdown is for serious GS2 Mains aspirants, especially anyone preparing for 2026 who wants to see how a topper reads a real paper. It suits candidates who keep "knowing" topics yet struggle to convert that knowledge into points under exam pressure, and who want a concrete demonstration of PYQ-driven preparation.

The strongest signal from the GS2 2025 paper is that previous year questions, plus basic sources, common sense and disciplined answer writing, are enough. To act on it, brainstorm each of these 20 questions and write full answers as part of regular daily answer writing practice, study the method for writing Mains answers, and test yourself under timed conditions with a structured Mains test series. For more paper breakdowns and strategy, explore the rest of the blog.