NEET PG: Paper Pattern & Weightage Analysis

Subject-wise and topic-wise patterns from the last four NEET PG exams (2023–2026), compiled as the input for building a study schedule.

Compiled 2 October 2026 · Covers NEET PG 2023, 2024, 2025 and 2026 (held 30 Aug 2026)

1. Read this first: data availability

The actual question papers are not publicly available

NBEMS does not release NEET PG question papers. Everything in circulation (“previous year papers”, PYQ banks) is memory-based recall, reconstructed by candidates and coaching institutes after the exam. These recalls are incomplete, and wording and options may differ from the real paper.
NBEMS no longer publishes a subject-wise blueprint. The only official subject distribution is the old table from when the paper had 300 questions. NBEMS dropped it from the bulletin after the move to 200 and then 180 questions. The current 2026 paper has 5 time-locked sections, and these sections are not organised by subject.
Different sources give conflicting subject counts. Several websites still present the old 300-question table as the “2023/2024/2025 distribution”, which is wrong. One 2025 recall table adds up to about 195 questions because it counts Dermatology twice. The large coaching platforms (PrepLadder, Marrow, DAMS) publish survey results, such as “84% of students said OBG was high-weightage”, but they don't publish exact counts.
For 2026, NBEMS said it would release answer keys from 1 October 2026, following a Supreme Court direction. These keys contain question IDs and correct options, not the question text, so they won't help with topic analysis.

What this report does instead: it takes the official baseline, scales it to 180 questions and adjusts it using recall evidence from 2023–2026. Every figure is labelled with its confidence level. Treat the numbers as planning estimates (±2–4 questions per subject), not exact counts.

2. How the exam format changed

Year	Exam date	Questions	Max marks	Structure	Reported character
2023	5 Mar 2023	200	800	Single CBT, 3.5 h	Moderate. Mix of one-liners and clinical questions
2024	11 Aug 2024	200	800	Two shifts, 5 timed sections (first year of time-locked sections)	Moderate to difficult. About 47% clinical, 51% conceptual, 2% one-liners (survey)
2025	3 Aug 2025	200	800	Single shift (Supreme Court order), 5 timed sections	Moderate. About 49% clinical, 41% conceptual, 10% one-liners. Many image-based questions reported
2026	30 Aug 2026	180	720	Single shift, 5 sections × 36 Q × 42 min, no going back to a section once it closes	Moderate to tough. Long clinical stems, highly integrated, very few images, no video questions, strong PYQ flavour

Marking has stayed at +4 / −1 / 0. With 180 questions, each question is now worth about 0.56% of the paper. A 5-question subject is worth roughly 2.8%.

3. Subject-wise weightage

3a. The official baseline (the last NBEMS blueprint)

This was the official distribution when the paper had 300 questions. Coaching analyses still use its proportions as the starting point. The table also shows the same proportions scaled to today's 180 questions.

Group	Subject	Official (of 300)	Share	Scaled to 180
Pre-clinical	Anatomy	17	5.7%	10
	Physiology	17	5.7%	10
	Biochemistry	16	5.3%	10
Para-clinical	Pathology	25	8.3%	15
	Pharmacology	20	6.7%	12
	Microbiology	20	6.7%	12
	PSM / Community Medicine	25	8.3%	15
	Forensic Medicine	10	3.3%	6
Clinical	Medicine + Dermatology + Psychiatry	45	15%	27
	Surgery + Ortho + Anaesthesia + Radiology	45	15%	27
	Obstetrics & Gynaecology	30	10%	18
	Paediatrics	10	3.3%	6
	ENT	10	3.3%	6
	Ophthalmology	10	3.3%	6

Confidence: high for the proportions, which are official. Low as a statement of what was actually asked in any one year.

3b. Working estimate for the 180-question paper

This adjusts the baseline using recall evidence from 2023–2026. Recalls show the clinical subjects growing (especially OBG and Medicine), PSM staying consistently heavy, and the pre-clinical subjects, especially Physiology, shrinking. The dark bar is the planning midpoint. The light band is the plausible range in a given year.

Estimated questions per subject (out of 180)

Planning estimate, not official. Hover over a row for details.

Midpoint estimatePlausible range

Subject	Range	Midpoint	Share	Trend 2023→2026	Priority

3c. How the big groups break down

Group (as NBEMS bundles it)	Inside the group	Est. range
Medicine + Derm + Psych (26–32)	General Medicine	20–24
	Dermatology & Venereology	3–5
	Psychiatry	3–5
Surgery + allied (24–30)	General Surgery (incl. urology, vascular, neuro, paediatric surgery)	15–19
	Orthopaedics	4–6
	Anaesthesia	3–4
	Radiology	3–4

Data point from one recall source (NEET PG 2025): Medicine 30, OBG 20, Surgery 18, PSM 16, Biochemistry 15, Pharmacology 15, Pathology 12, Microbiology 12, Anatomy 9, Dermatology 8, FMT 6, Physiology 5, Paediatrics 5, Ortho 5, Ophthalmology 5, ENT 5, Radiology 5, Anaesthesia 4.

Shown only for reference. The total is about 195, with Dermatology double-counted, and it comes from a single recall set. It agrees with the trend (clinical subjects heavy, Physiology light) but is too unreliable to use on its own.

4. How questions are being asked

Dimension	2024	2025	2026	Direction
Clinical vignettes	~47%	~49% (some recall sets: ~70%)	Dominant (76% of surveyed students called it the main type)	↑ Rising every year
One-liners / pure recall	~2%	~10%	Minimal	↓ Falling
Image-based	Many (Path, Micro, Derm)	High (some estimates ~50%)	Very few	↕ Volatile. Still prepare for images
Integrated / multi-subject	High	High	Very high	↑ Rising
PYQ repeats / PYQ-style	Moderate	~5–20 repeats (up to ~30% reported by some)	Heavy PYQ emphasis noted	→ Consistently important
Difficulty (student survey)	Moderate–difficult	82% said moderate	55% moderate, 40% difficult	↑ Harder, longer stems
Time pressure	—	16% said time was not enough	32% said time was not enough	↑ Speed now matters

Percentages are from student surveys run by coaching platforms (PrepLadder, Cerebellum and others). They are shares of respondents, not exact shares of questions.

What this implies:

Questions now typically go clinical scenario → diagnosis → next step, drug or mechanism. For example, a Pharmacology fact is tested inside a Medicine case. Study each topic across subjects, not as a standalone fact.
“Next best step” and management questions dominate Medicine, Surgery and OBG. Learn management algorithms and current guidelines, not just features of each disease.
Long stems combined with locked 42-minute sections mean timed mock tests have to be part of the schedule.

5. Topic-wise high-yield areas, by subject

For each subject, chapters are ranked by how often they have appeared across recall-based PYQs over the years. “Recalled 2025/2026” lists topics that students specifically reported from those exams. Click a subject to expand it.

Confidence: the chapter priorities are a consensus of recall-based PYQ analyses and are fairly stable from year to year. The recalled topics are reported examples, not complete lists.

6. What this means for the schedule

Tier	Subjects	Est. share of paper	Schedule implication
Tier 1	Medicine, Surgery, OBG, PSM, Pathology	~54%	Most of the time, two or more revision cycles, and the bulk of question practice. Pathology underpins Medicine, so pair the two.
Tier 2	Pharmacology, Microbiology, Anatomy, Biochemistry	~23%	Full coverage, but focus on high-yield material. Learn Pharmacology and Microbiology alongside the clinical systems.
Tier 3	Physiology, Paediatrics, FMT, ENT, Ophthalmology, Derm, Psych, Ortho, Anaesthesia, Radiology	~23%	Short, focused blocks. These are small subjects, but together they are almost a quarter of the paper and give the best marks per hour. Never skip them.

Yield per study hour matters as much as raw weightage. FMT, Ophthalmology, ENT, Psychiatry and Dermatology each take only a few days to finish and are worth 3–7 questions. Medicine is worth the most questions but has the lowest return per hour.
PYQs are non-negotiable. Each year shows repeats or near-repeats, and 2026 had a strong PYQ flavour. The schedule should include at least one full pass through subject-wise PYQs.
Recent changes are tested. Expect questions on updated national programmes (for example, NTEP regimens) and updated laws. Since July 2024, FMT questions use the new criminal laws (BNS/BNSS/BSA) in place of IPC/CrPC.
Practise timed sections. In the 2026 format, each 36-question section has 42 minutes (70 seconds per question) and you can't go back. Mock tests should use the same structure.
Image practice is cheap insurance. 2026 had few image questions, but 2024–25 had many. Keep short image-bank sessions for Pathology, Microbiology, Dermatology, Ophthalmology, Radiology, Anatomy and FMT.

7. Gaps & how to close them

Gap	Why it matters	How to close it
No official question papers	Exact topic counts per year can't be verified	Not fixable. NBEMS policy and court rulings keep the papers confidential.
No reliable, complete subject counts for 2023–2026	The weightage figures here are estimates (±2–4 per subject)	Collect the full recall sets (PrepLadder, Marrow, DAMS and similar, usually in their apps) for 2023–2026. Each question can then be tagged by subject and topic and counted, which would give real per-year counts.
2026 recall data is still thin (the exam was 5 weeks ago)	The newest trends rest on surveys and partial recall lists	Revisit in a few weeks, when the complete 2026 recall compilations are out.
Some free “PYQ” PDFs are low quality or generated	Generic questions can distort topic counts	Use only recall sets from reputable sources, and cross-check topics across at least two sources.