What 447 Ivy League Decisions Say About Who Gets In
A tally of 447 self-reported 2026 Ivy League decisions points to a GPA plateau at 3.9, a big ACT 36 gap, research with real output, and no early-round advantage for unhooked applicants.
No single Ivy gets enough public results posts to say much about it alone. Harvard, for example, admitted only 3 of the 54 applicants in one tally. Pool all eight Ivies together, though, and you get 447 decisions from 113 applicants who posted their full 2026 results on r/collegeresults. That’s enough to show some useful patterns.
First, how to read the numbers
The sample’s admit rate is about 17%. The Ivies’ real average is around 5%. The difference comes from who posts: students with strong profiles and good news are more likely to share. So 17% is not anyone’s odds. What matters is the gap between groups: applicants who had something versus applicants who didn’t.
Recruited athletes and transfers are excluded, every number is self-reported, and someone who applied to six Ivies counts six times. Some groups are small, so treat the patterns as signals, not laws.
GPA: the line is 3.9, not 4.0
- Below 3.9 unweighted: admitted in about 7% of decisions
- 3.9 to 3.99: about 24%
- 4.0: about 20%
Under 3.9, the rate fell sharply. Above it, a perfect 4.0 did no better than a 3.9-something. Once you clear the bar, the decimals stop mattering.
Test scores: the ACT stood out
Among applicants who sent scores, ACT 36 scorers were admitted in 30% of their Ivy decisions, compared with about 7% for scores below 36. That was the largest gap in the whole analysis. The SAT showed no gap like it (about 21% at 1550+ versus 17% below). Fewer than 40 ACT takers were in the sample, so this gap could shrink as more results come in.
Course load mattered too. Applicants admitted somewhere in the Ivies had a median of 14 AP/IB courses, compared with 10 for those admitted nowhere.
Activities: research with real output
The biggest activity gap came from a research publication or paper award: 33% versus 14%. Research of any kind, named scholarships, and debate or Model UN also lined up with higher admit rates.
The surprise was technical work. Coding projects (9% versus 20%) and robotics or engineering teams (8% versus 20%) came with lower odds. That doesn’t mean they hurt. It’s more likely that so many Ivy applicants list them that they don’t stand out. Varsity sports and community service made almost no difference either way.
Applying early didn’t help here
Early applications were admitted 14% of the time, compared with 18.5% in regular decision. Most early applicants were deferred, and only 2 of 26 deferred applicants were admitted in the spring. That conflicts with what schools like Brown report, but this sample is almost entirely unhooked applicants. For them, the early round wasn’t a shortcut.
Key takeaways
- Get above 3.9, then stop worrying about the decimals.
- A top ACT score was the strongest academic signal in this data.
- Research that produced something real set applicants apart. Common technical projects didn’t.
- Don’t count on early rounds or deferrals to change your odds.
- Ivy outcomes cluster. Applicants admitted to one Ivy often got into others too, which suggests the same profile tends to succeed (or not) across all of them.
For the full tables, school-by-school admit and waitlist numbers, and notes on legacy and first-gen applicants, read CollegeIntel’s analysis of Ivy League accepted vs. rejected applicants. It’s a useful reality check for anyone building an Ivy list this fall.