GATE DA Marks vs Score vs Rank: Official Formula and Calculator

The DA trend is the opposite of what many candidates assume. As the paper grew from 39,000 to 69,000 candidates, the average mark fell from 23.4 to 15.5 — whether because the paper got harder, the field got broader, or both, the numbers alone cannot tell. The effect on you is clear, though: the same marks earned a higher score in 2026 than in 2024. Scores can also be negative: the lowest 2026 DA mark was −13.67, and the formula gives such a mark a negative score.

The same numbers as tables, for readers on a simplified view:

Computed with the official formula and each year's official Mq and Mt, capped at 1000 (no floor — very low marks give negative scores). As a check, it reproduces the official 2026 OBC-NCL/EWS qualifying score of 317 and the 2025 figure of 316.

Each cell is the number of candidates at or above that mark, added up from the official 10-mark distribution table in each report. Scoring exactly that mark, your rank would have been about that number or better — it is an upper bound. Our arithmetic on official data.

From table 10 of each official report (starting and ending rank of qualified candidates in each score band). Rank counts everyone who appeared with a higher score, which is why ranks run beyond the number who qualified.

Qualifying marks and scores as printed in the official reports. Qualifying only means your score is valid for three years; almost every M.Tech programme at an IIT or IISc sets its own, much higher, cut-off, and some PSUs and companies that recruit through GATE set theirs higher still.

Most GATE papers have a decade of history, so candidates can see how their qualifying marks drift. DA has three data points, and they point in one direction for a reason that may not last: in its first three years the paper roughly doubled in size, pulling in candidates from computer science, electronics, mathematics and statistics who had never taken it before. A broader field lowers the average, and the average drives the qualifying mark.

Two things could reverse that. If DA's growth slows, the field will settle into candidates who prepared specifically for it, and the average may rise. And a paper-setting committee that finds 2026 too hard can set an easier 2027 — every organising institute is different. The safe planning assumption is therefore not "the qualifying mark keeps falling" but "a score of 600 will need somewhere between 47 and 55 marks", the range the three years actually span.

Compare DA with Computer Science, its nearest neighbour. CS had 211,020 candidates in 2026 across two normalised sessions and a qualifying mark of 30; DA had 69,242 in a single session and 26.4. The shares who qualified were close (19.3% for CS, 18.6% for DA), but the top end was thinner in DA: only 0.05% of DA candidates scored 70 or more, against 0.14% in CS.

On the 2024 field (Mq 37.1, Mt 76.99): 350 + 550 × 2.9 / 39.89 ≈ 390, and about 5,160 candidates scored 40 or more. On the 2025 field (Mq 29, Mt 75.89): 350 + 550 × 11 / 46.89 ≈ 479, with about 2,570 at 40+. On the 2026 field (Mq 26.4, Mt 72.01): 350 + 550 × 13.6 / 45.61 ≈ 514, with about 2,410 at 40+.

So the same 40 marks moved from a score near 390 to one above 500, and from roughly rank 5,000 to roughly rank 2,400. That is the strongest argument against setting your target as a fixed mark: set it as a score or rank, and use the latest year to translate.