.png)
The best law schools for splitters, ranked by the one number that actually predicts splitter tolerance: the gap between a school's GPA median and its GPA 25th percentile. WashU leads at 0.38, meaning a quarter of its class enrolled at 3.58 or below against a 3.96 median. Here are all 31 schools.
The wider the gap, the further below its own median a school routinely goes on GPA. That is splitter tolerance expressed as a disclosed number rather than as folklore.
#SchoolGPA 25thGPA medianGapLSAT medianAcceptance1WashU3.583.960.3817519.0%2Texas A&M3.674.000.3316912.1%3GW3.553.860.3116827.2%4Ohio State3.643.910.2716824.6%5Minnesota3.623.880.2617126.7%6UCLA3.733.950.2217112.1%7Georgia3.703.920.2216912.7%8Northwestern3.763.960.2017312.3%9Wisconsin3.613.810.2016719.1%10Wake Forest3.603.790.1916624.6%11Penn3.773.950.181738.1%12USC3.733.910.1816911.2%13Georgetown3.753.930.1817115.8%14BU3.713.880.1717012.1%15Cornell3.753.920.1717318.2%16Boston College3.673.830.161688.5%17UVA3.833.990.1617310.2%18Michigan3.743.880.141718.6%19Vanderbilt3.773.910.1417013.3%20Texas3.753.890.1417214.0%21BYU3.823.950.1317022.8%22UNC3.783.890.1116811.2%23NYU3.813.920.1117213.4%24Notre Dame3.783.890.1117016.1%25UChicago3.873.970.101749.7%26Stanford3.873.960.091736.1%27Duke3.833.910.0817112.9%28Berkeley3.843.920.0817014.8%29Harvard3.893.960.071749.2%30Columbia3.853.920.0717311.8%31Yale3.903.960.061744.1%
Sources: each school's 2025 ABA 509 report and class of 2025 ABA Employment Summary, verified July 26, 2026. The gap column is our arithmetic from the two disclosed percentiles.
A splitter is an applicant whose LSAT sits at or above a school's median while their GPA sits at or below its 25th percentile. The reverse splitter is the mirror image, and it is a harder position, which our reverse splitter guide covers.
Splitters exist because schools report both medians to the ABA every December and the two numbers are not equally easy to buy. An applicant delivering a high LSAT hands the school the scarcer number, so schools accept a GPA hit to get it.
That trade is visible in every row above. A school posting a 0.30 gap is disclosing that a quarter of its enrolled students arrived at least three tenths below the class midpoint, and those students almost all had LSATs doing the work.
Read the top of the table rather than any list you have seen elsewhere. WashU at 0.38, Georgia at 0.22, Texas A&M at 0.33 and Wake Forest at 0.19 all tolerate GPA spread that the very top of the market does not.
The pattern is structural. Schools competing hardest on LSAT median, which is the number rankings and applicants watch most closely, buy it with money and with GPA flexibility.
Compare the extremes. Columbia's gap is 0.07 and Yale's is 0.06, so both effectively require a GPA within a tenth of their median; WashU's 0.38 and Texas A&M's 0.33 leave real room for a strong-LSAT file with a damaged transcript.
One caution about the very top of the table. A wide gap tells you a school has admitted low-GPA students, not that it will admit you, and it says nothing about how much scholarship money follows.
Yes, and the disclosed gaps show which T14 schools are the realistic targets. Penn at 0.18, Cornell at 0.17, Michigan at 0.14 and Northwestern at 0.20 run materially wider GPA spreads than Yale, Harvard, Columbia or Stanford.
The working profile at those schools is an LSAT at or above the median paired with a GPA in the 3.6 to 3.8 range. At Northwestern that means a 173 over a 3.76, which is exactly the disclosed 25th percentile.
Below that, the math gets hard fast. A 3.4 against Columbia's 3.85 twenty-fifth percentile is four and a half tenths under the floor, and no essay closes that distance.
Lead with the LSAT and stop apologizing for the GPA. Your file's argument is that you can do the work, and the LSAT is the number that makes it.
Write a GPA addendum only if there is a documentable story: an illness, a work schedule, a family situation, or a clear upward trend across semesters. Under one page, factual, no pleading, closing on what changed.
Apply early, because splitters are exactly the files that benefit from being read when seats are plentiful. And apply broadly across the gap column rather than clustering at one tier, since splitter outcomes vary more school to school than balanced applicants' do.
Getting in is half the problem. Merit aid tracks how far your numbers sit above a school's medians, and a splitter is above on one and below on the other.
That usually produces a smaller award than a balanced applicant with the same index would get. The schools with the widest gaps often compensate: WashU's median grant is $48,000 to 94 percent of students, and Wake Forest's is $40,000 to 76 percent.
Run the aid math before you fall in love with a reach. Our negotiation playbook covers how to move the number once offers arrive.
WashU posts the widest GPA spread in this set at 0.38, with a 3.58 twenty-fifth percentile beneath a 3.96 median, alongside a $48,000 median grant reaching 94 percent of students. It is the clearest splitter target here on both admission and money.
Texas A&M at 0.33 is the second widest, and its 509 prints a 4.00 median GPA against a 3.67 floor. Georgia at 0.22 pairs a 3.70 floor with a 159 LSAT twenty-fifth percentile, which means it flexes on both numbers rather than just one.
Northwestern at 0.20 is the most useful row for T14-focused splitters, with a 3.76 GPA floor under a 3.96 median and a $40,000 median grant to 76 percent of students. Penn at 0.18 and Cornell at 0.17 follow.
Read those five against the acceptance column before getting excited. A wide gap at a school admitting under 10 percent still means a hard admission; the gap tells you which of your two numbers will sink you, not whether you get in.
The LSAT does the arguing, so lead with it and do not spend your personal statement apologizing for the transcript. Committees read the numbers first and form a position before your writing arrives.
Add a GPA addendum only where a documentable event explains the record: illness, a work schedule, a family obligation, or a first year that improved steadily afterward. Under one page, factual, chronological, closing on what changed rather than on how much you grew.
Get one recommender to speak directly to your academic capacity, because the GPA raises exactly that question and a supervisor letter about work ethic does not answer it. And submit early, since splitters benefit most from being read when seats are plentiful.
Every number in the table comes from a primary disclosure: each school's 2025 ABA Standard 509 Information Report for admissions, cost and aid, and its class of 2025 ABA Employment Summary for outcomes. Both are filed annually and both are free to read.
We did not use rankings, surveys or reputation. Where a figure required arithmetic, such as a percentage or a net cost, we state that it is our arithmetic and give the inputs so you can check it.
Where we could not verify something, the table says so rather than filling the gap with an estimate. That is why some rows read not published or not verified instead of carrying a number that would look more complete and mean less.
It cannot tell you your outcome. Every figure here describes a class of several hundred people, and your position inside that distribution is set by first-year grades that have not happened yet.
It cannot price your offer, because aid depends on where your numbers sit against each school's medians rather than on the median award. Half of every school's grant recipients got less than the figure shown.
And it cannot capture fit, culture, or whether you will do your best work in that building for three years. Those are discoverable only by visiting and asking current students specific questions, which is worth more than any table on this site.
Applying only to wide-gap schools. The gap tells you which number will sink you, not that admission is easy, and a list built entirely from the top four rows is still a list of hard admissions.
Writing the addendum as an argument. A committee wants a fact pattern with a date and a change, not a case for why the GPA misrepresents you. The second reads as litigation and invites scrutiny.
Waiting for a higher LSAT that never arrives. If your score is already above a school's median, additional points buy scholarship rather than admission, and a cycle spent chasing them costs a year.
Ranked by the gap between GPA median and GPA 25th percentile, the widest are WashU at 0.38, Texas A&M at 0.33, Georgia at 0.22 and Northwestern at 0.20. A wide gap means the school routinely enrolls students well below its own GPA midpoint.
Yes. Penn, Cornell, Michigan and Northwestern all run GPA spreads of 0.14 or wider, so a high LSAT over a 3.6 to 3.8 GPA is a live file. Yale, Harvard, Columbia and Stanford run spreads under 0.10 and are much harder.
An applicant whose LSAT is at or above a school's median while their GPA is at or below its 25th percentile. Schools admit splitters because they report both medians to the ABA and the LSAT is the harder number to acquire.
Only if there is a documentable reason: illness, work obligations, a family situation, or a clear upward grade trend. An addendum that simply regrets a low GPA draws attention to the weakness without explaining it.
Often, because merit aid tracks how far above a school's medians you sit and a splitter is below on one of the two. Schools with wide GPA spreads frequently discount deeply, which partly offsets it.
Written by Ali, Georgetown Law, founder of Lovare Institut.
July 30, 2026
July 30, 2026