Every number in this post is computed directly from MCC's own official PG Counselling 2025 allotment result PDFs -- Round 3 (1,555 pages) and the Stray Vacancy Round (130 pages) -- using the same pipeline that powers the Predictor tab on NEET-PG Help. Nothing here is a third-party estimate.
The headline numbers
| Round 3 | Stray Vacancy Round | |
|---|---|---|
| Institutes with at least one seat allotted | 1,499 | 743 |
| Distinct specialties | 97 | 85 |
| Total seats counted | 18,673 | 2,798 |
What that actually means
The Stray Vacancy Round is, as the name says, a mop-up round -- it only allots the seats still vacant after Round 3 (and any subsequent state-level processes) conclude. So a much smaller round is expected. What's more interesting is which institutes show up: of the 743 institutes with a seat allotted in the Stray round, all but one also had a seat allotted in Round 3 -- but 757 institutes that had a Round 3 seat had nothing left to offer by the Stray round. That's a real signal of which colleges and specialties fill up early versus which ones still have seats months into the cycle.
Category-wise split
| Category | Round 3 seats | Stray seats |
|---|---|---|
| Open | 5,987 | 1,363 |
| OBC | 2,209 | 446 |
| SC | 1,422 | 383 |
| EWS | 883 | 120 |
| ST | 777 | 171 |
The proportions hold up fairly consistently between the two rounds -- Open seats are roughly 32% of the total in both -- which suggests the Stray round isn't skewed toward any one category more than Round 3 was, at least at this aggregate level.
What this means if you're preparing for PG Counselling 2026
Don't compare a Stray-round closing rank directly to a Round 3 closing rank for the same seat and assume they tell you the same thing -- they come from different, non-overlapping pools of candidates at different points in the cycle, and our Predictor labels them separately for exactly this reason. If you want to see where a specific institute or specialty landed in either round, the Predictor tab lets you pick the round explicitly and shows you real reported ranks, not a blended estimate.