The Two Documents and What Each Contains
Within days of the last shift of a session, NTA publishes two things together. The provisional answer key lists the correct answer for every question in every shift. Your response sheet is a record of your own attempt — the questions as you saw them, the option you selected or the value you entered, and your question identifiers.
Neither is your result. The result comes later, carries a percentile rather than raw marks, and is calculated after the final answer key is settled. What these two documents give you is the ability to compute your raw score yourself, precisely, before anyone tells you anything.
Download Both Immediately
The challenge window is short — typically two to three days — and the response sheet is not available indefinitely afterwards. Download both the day they appear, save them locally, and take screenshots as a backup. Candidates who intend to look at it over the weekend regularly find the window closed.
Note also that question numbering differs between candidates. The exam shuffles question order, so your response sheet must be matched against the answer key using the question identifiers rather than by position. Comparing your sheet against a friend's is meaningless for the same reason.
Calculating Your Raw Score
The arithmetic is simple and worth doing carefully rather than quickly. Award four marks for every correct answer. Deduct one mark for every incorrect answer, in both Section A and Section B under the current marking scheme. Award nothing and deduct nothing for unattempted questions. Sum across all three subjects for a total out of 300.
Do this subject by subject rather than as one total, because the subject split is where the diagnostic value lives. A total of 190 built from 80 in Chemistry and 45 in Physics tells a completely different story from the same total built evenly, and it points at a completely different plan for the weeks before the next session. The marking rules in full are in the exam pattern guide.
The Analysis Most Candidates Skip
Once you have the score, resist the urge to stop. The response sheet is the most accurate diagnostic document you will ever hold about your own exam performance, because unlike a mock it records what you did when it genuinely counted. Spend an hour on it.
- Count your attempts per subject. Were you attempting too few out of caution, or too many out of optimism?
- Calculate your accuracy — correct answers as a share of attempted. Below roughly eighty percent usually indicates a guessing problem rather than a knowledge problem.
- Count how many marks you lost to negative marking alone. This number is often startling and is almost always reducible.
- Identify unattempted questions you could actually have solved. These are triage failures, and they are the cheapest marks to recover.
- Check your Section B usage: did you attempt five per subject, and did you pick the five most tractable of the ten?
The pattern that emerges is your instruction set for the next ten weeks. A candidate losing thirty marks to negatives needs discipline, not more chapters. A candidate with high accuracy and low attempts needs coverage and confidence. The two problems look identical on a scorecard and require opposite responses — the framework is set out in the attempt strategy guide.
One comparison worth making while you have both documents open is between this attempt and your recent mock attempts. If your exam accuracy is markedly lower than your mock accuracy, the gap is almost never knowledge — it is the effect of exam-hall pressure on question selection, and it is trainable by attempting mocks under stricter conditions rather than by studying more. If your exam accuracy matches your mocks, your practice conditions were honest and the score is a fair reading of where you stand. Very few candidates run this comparison, and it answers a question that otherwise gets guessed at for weeks.
When a Challenge Is Worth Filing
NTA allows candidates to challenge the provisional answer key within the stated window, on payment of a non-refundable fee per question challenged. Challenges are reviewed by a subject expert panel, and the outcome is not communicated individually — it appears as a change in the final answer key, if the challenge succeeds.
The test for whether to file is not whether you disagree with the key but whether you can support the disagreement with a citation. A challenge accompanied by a reference from NCERT or a standard text has a genuine chance. A challenge that amounts to a differing opinion has effectively none, and the fee is gone either way.
Do it when the question is one you attempted and got marked wrong, the key looks genuinely incorrect or ambiguous, and you can cite a source. Do not do it because a coaching institute posted a different answer, or because several people online agree with you. Where a question is genuinely defective, NTA has historically dropped it and awarded full marks to all candidates, which means a legitimate error usually gets caught with or without your fee.
What the Final Answer Key Changes
After the challenge window closes, NTA publishes the final answer key alongside the result. Some answers may have changed, and some questions may have been dropped entirely with full marks awarded to everyone. Your result is computed from the final key, not the provisional one.
This means your self-calculated raw score can shift by a few marks in either direction, which is normal and not worth agonising over. What it does not change is the diagnostic picture — your accuracy, your attempt count and your subject balance are unaffected by two questions being dropped, and those are the numbers you should be acting on.
From Raw Score to Percentile
The most common mistake at this stage is converting your raw score into an expected percentile using last cycle's chart and then treating the result as fact. The mapping is rebuilt each session from that session's candidate pool and paper difficulty, so a previous chart gives you a band, not a number.
Hold the estimate loosely and stop refreshing prediction tools. The mechanism that turns your marks into a percentile is explained in the normalisation guide, and the reason the curve behaves so unintuitively at the top is in marks versus percentile. Neither will make the wait shorter, but both will stop you from drawing false conclusions from it.
Using the Ten Weeks Between Sessions
If you are appearing in Session 2, the answer key arrives at almost exactly the right moment: early enough that there is time to act, specific enough to tell you what to act on. This is the single most valuable feedback loop in the entire cycle, and most candidates waste it by spending the window on percentile speculation.
Convert the analysis into a plan the same week. Two or three named weaknesses, a target for negative marks, and a fixed schedule of full-length papers under real conditions — attempted on the free JEE Main mock test so the interface stays rehearsed as well as the syllabus. The specific structure for a ten-week repair window is in the Session 1 versus Session 2 guide, and the compressed schedule in the 90-day study plan.