Williams College: Inside the Numbers

July 9, 2026

Williams runs a peculiar economy. A college of 2,100 undergraduates in a small Berkshires town competes annually for cross-admits against Harvard, Yale, and Stanford, sits at or near the top of the U.S. News liberal arts ranking year after year, and in 2022 became the first college in the country to convert every loan and work-study expectation in its financial aid packages into outright grants. The five admission cycles from Fall 2021 to Fall 2025 capture Williams absorbing three external shocks (the pandemic-era surge in applications, the Supreme Court鈥檚 2023 SFFA decision on race-conscious admissions, and the broader unwinding of test-optional policies) while making one large internal commitment to all-grant aid.

What follows walks through the five most recent Common Data Sets, from 2021鈥22 through 2025鈥26, with an eye toward what families considering Williams, and the counselors advising them, should take from the data. Where it sharpens the picture, I have pulled in publicly reported figures the CDS does not capture, including the regular decision and waitlist breakouts Williams releases each March, and the demographic data the college shared with the Williams Record after the Court ruling.

1. Applications, admits, and yield

Williams鈥 application volume reached its pandemic-era peak in Fall 2022, then dipped sharply for Fall 2023 before rebounding to a new high in Fall 2025. The Fall 2023 dip (11,465 applications, down from 15,321 the prior year) is the single largest year-over-year swing in the window and worth understanding. It coincided with two changes: Williams reinstated a required short supplemental essay (one cycle after making writing supplements optional for Fall 2022), and the regular decision pool contracted, dropping by more than a quarter year over year per the Williams Record. The all-grant aid announcement in April 2022, which preceded the Fall 2023 cycle, did not pull in additional applications in year one; the broader pool came back the following year.

Applications & admit rate 2021鈥22 2022鈥23 2023鈥24 2024鈥25 2025鈥26
Applications 12,452 15,321 11,465 15,411 15,520
Admits 1,099 1,302 1,145 1,272 1,336
Admit rate 8.8% 8.5% 10.0% 8.3% 8.6%
Enrolled 574 577 541 547 567
Yield 52.2% 44.3% 47.2% 43.0% 42.4%

Source: Williams College Common Data Set, C1, 2021鈥22 through 2025鈥26. Admit and yield rates calculated from raw counts.

Over the five cycles, the admit rate has settled into a tight range between roughly 8 and 10 percent. The Fall 2024 cycle (Class of 2028) recorded what the Williams Record called a record-low overall acceptance rate (their calculation of 7.5 percent uses the initial admits released in March; the 8.25 percent figure in the CDS reflects later waitlist activity, including 113 students admitted from the waitlist that year). The most recent cycle ticked up slightly to 8.6 percent at CDS reporting, in part because Williams raised its target class size from roughly 550 to 560 for the Class of 2029.

Yield is the quieter figure. Williams鈥 yield ran above 50 percent in Fall 2021, then settled into the low-to-mid 40s through Fall 2025. That is not unusual for highly selective liberal arts colleges; the post-pandemic market has been hard on yield at peer institutions for similar reasons, as cross-admits between selective universities and selective LACs have widened and applicants increasingly hold offers from multiple institutions before deciding. A yield in the mid-40s remains strong by national standards and reflects the cross-admit market Williams competes in.

2. The Early Decision route

Early Decision applications climbed every year through Fall 2024, when the ED pool hit a record 1,067 applications. The ED admit rate compressed accordingly, dropping from 33.3 percent for Fall 2021 to a record-low 23.3 percent for Fall 2024 before bouncing back to 26.6 percent for Fall 2025, when ED volume eased to 964 applications.

Early Decision 2021鈥22 2022鈥23 2023鈥24 2024鈥25 2025鈥26
ED applications 715 814 943 1,067 964
ED admits 238 254 255 249 256
ED admit rate 33.3% 31.2% 27.0% 23.3% 26.6%

Source: Williams CDS, C21.

The ED admit rate has run roughly two to three times the regular decision rate every year, and that gap has grown. For Fall 2024, the regular decision admit rate was about 6.4 percent (roughly 1,159 RD admits among about 14,100 RD applications) against an ED rate of 23.3 percent. For Fall 2025, the RD rate was about 7.3 percent. That is not a thumb on the scale for ED applicants as such; it reflects the reality that ED applicants self-select on commitment to Williams, often have stronger fit signals on the file, and lock in yield for the office at the moment of admission. Williams typically fills roughly 45 percent of its target class through ED, and that share has held steady even as the math underneath it has shifted.

For applicants ranking Williams as a clear first choice, the math favors ED. For those still triangulating between Williams and other top LACs or Ivy-tier universities, the binding commitment is the real constraint. There is no early action or restrictive early action option, and Williams stays conservative on deferred admits from the ED round, typically admitting 14 to 30 deferred candidates in regular decision, per recent Record reporting.

3. Testing, scores, and what truly test-optional means

Williams has been test-optional since the pandemic and, as of the 2025鈥26 CDS, remains so. The college鈥檚 C7 listing for standardized test scores shifted from 鈥淚mportant鈥 in the 2021鈥22 and 2022鈥23 CDS to 鈥淐onsidered鈥 from 2023鈥24 onward, consistent with how the office describes its review: scores supplement the file but are not required, and the office tells applicants that the policy is 鈥渢ruly鈥 optional rather than test-optional-but-strongly-recommended.

SAT (submitters) 2021鈥22 2022鈥23 2023鈥24 2024鈥25 2025鈥26
SAT 25th 1470 1490 1480 1500 1490
SAT median 1510 1520 1520 1535 1520
SAT 75th 1550 1550 1550 1560 1550
ACT (submitters) 2021鈥22 2022鈥23 2023鈥24 2024鈥25 2025鈥26
ACT 25th 33 34 33 34 33
ACT median 34 35 34 34 34
ACT 75th 35 35 35 35 35

Source: Williams CDS, C9. Figures reflect students who submitted scores; the submitting share has ranged from about half to about 70 percent of the enrolled class across the window.

The middle 50 percent SAT range has hovered between roughly 1480 and 1560 across the five cycles. The 25th percentile has crept upward, from 1470 to a peak of 1500 for Fall 2024, settling at 1490 most recently, while the 75th has stayed pinned at or just above 1550. ACT composites say the same thing, 33 to 35 every year, with a 34 to 35 middle 50 for Fall 2022 and Fall 2024. The practical takeaway: an applicant with scores at or above the 75th percentile (1550 SAT, 35 ACT) is almost always helped by submitting them. Below the 25th, scores often do not add to the file, and most applicants in that range correctly choose not to submit.

One pattern is worth understanding. In Fall 2021, roughly 71 percent of enrolled first-years submitted SAT or ACT scores (43 percent SAT, 28 percent ACT). By Fall 2024, that combined share had dropped to about 52 percent (35 percent SAT, 17 percent ACT), yet Williams鈥 middle 50 ranges held steady. That tells you the population of submitters is increasingly self-selecting at the top end, and the middle 50 on the CDS reflects a narrower, more accomplished slice of the class than it did five years ago. A 1500 in Fall 2021 was roughly the 25th percentile of submitters when most students submitted; a 1500 in Fall 2025 is the 25th percentile of a much smaller, top-end-skewed group.

4. What Williams says matters in the file

Williams鈥 Common Data Set C7 listing of admission factors has been stable across the five cycles. The academic factors rated 鈥淰ery Important鈥 every year are rigor of secondary school record, class rank, academic GPA, and recommendations. Application essays have moved between 鈥淰ery Important鈥 and 鈥淚mportant鈥 depending on the cycle and the CDS template. Standardized test scores moved from 鈥淚mportant鈥 (2021鈥22 and 2022鈥23) to 鈥淐onsidered鈥 from 2023鈥24 onward, mirroring the entrenchment of the test-optional policy.

On the nonacademic side, character/personal qualities is the single nonacademic factor Williams has consistently rated 鈥淰ery Important.鈥 Talent/ability, extracurriculars, first-generation status, volunteer work, and work experience all sit at 鈥淚mportant.鈥 Alumni relation and geographical residence are 鈥淐onsidered.鈥 Two factors are consistently 鈥淣ot Considered鈥: interview (Williams does not offer evaluative interviews) and level of applicant鈥檚 interest.

That last one, Williams鈥 鈥淣ot Considered鈥 stance on level of applicant鈥檚 interest, deserves a careful reading. The college does not formally track who visited campus, opened an email, or attended an info session as part of the decision. But student-side contacts (information requests through the admissions website, attendance at virtual sessions, meetings with admissions reps at high school visits, joining QuestBridge or a fly-in program) still register in the applicant file, and they can sharpen the picture of an applicant鈥檚 seriousness even where they are not formally weighted. A campus visit is not essential. Lower-cost engagement, especially for students far from Williamstown, signals real interest at minimal expense.

5. Class composition and the post-SFFA picture

Williams鈥 first-year demographics shifted modestly in the first two classes admitted after the Supreme Court鈥檚 June 2023 SFFA decision (the Class of 2028, enrolling in Fall 2024, and the Class of 2029, enrolling in Fall 2025). The most visible recent change is the drop in Black enrollment in the most recent class.

Class composition (B2) 2021鈥22 2022鈥23 2023鈥24 2024鈥25 2025鈥26
Hispanic / Latino 12.9% 15.8% 14.8% 15.5% 13.5%
Black 3.5% 7.1% 6.8% 7.3% 4.8%
White 52.6% 42.5% 45.3% 42.6% 43.2%
Asian 11.8% 13.7% 12.6% 11.5% 14.6%
Two or more races 7.1% 8.1% 7.9% 7.9% 6.2%
International 7.3% 9.0% 9.1% 11.0% 10.3%
Race unknown 4.7% 3.8% 3.3% 4.2% 7.5%

Source: Williams CDS, B2. Shares computed against B2 total first-time first-year counts; figures may vary slightly from press releases that use different denominators.

The Class of 2028 (Fall 2024, the first cohort admitted under the new legal framework) showed almost no demographic disruption, with Black share rising slightly to 7.3 percent and Hispanic share to 15.5 percent compared to the prior class. Williams attributed that continuity to longstanding race-neutral recruitment efforts and to the recruiting impact of the all-grant aid policy. The Class of 2029 (Fall 2025) reads differently: Black share dropped from 7.3 to 4.8 percent, the lowest in the window, while the share of students who declined to disclose race or ethnicity rose to 7.5 percent. Hispanic share also moved down, to 13.5 percent, though more modestly. The international share has roughly doubled across the window, from 7.3 percent in Fall 2021 to about 10 percent in the most recent classes.

The Fall 2025 demographic data should be read carefully. A single-year movement is not a trend, the rising race-unknown share absorbs some of the apparent decline in named categories (students who in past years checked Black or Hispanic may now decline to identify), and the overall class size also expanded. That said, the Fall 2025 Black share is notably lower than the Fall 2022 through Fall 2024 baseline, and it is worth asking the admission office directly, on campus visits and in information sessions, how they are thinking about recruitment in the second post-SFFA cycle and beyond.

6. Cost is up. The aid program has gone further.

Williams鈥 comprehensive fee (tuition, required fees, food, and housing) reached $90,750 for the 2025鈥26 academic year, a 5.7 percent increase, and the full cost of attendance, once books, travel, and personal expenses are added, comes to about $95,960. That increase tracks with the broader cohort of highly selective private colleges. What sets Williams apart is not the sticker price; it is the aid program.

In April 2022, Williams became the first college in the country to convert every loan, campus-job, and summer-earnings expectation in its aid packages into outright grants. The program took effect for the 2022鈥23 academic year (the cohort entering as the Class of 2026) and applied to current students as well. The financial signature of that change is visible across the H1 and H2 sections of the CDS. Need-based loan dollars in the first-year aid package dropped from $825,584 in 2021鈥22 to zero in 2022鈥23 and have stayed at zero. Federal Work-Study disappeared from the package the same year. The share of the first-year class receiving need-based aid rose from 44 percent for Fall 2021 to roughly 56 to 58 percent in each of the four subsequent classes.

Need-based aid, first-years 2021鈥22 2022鈥23 2023鈥24 2024鈥25 2025鈥26
% on need-based aid 44.3% 57.9% 53.4% 55.9% 56.8%
Avg aid package $61,853 $65,134 $70,764 $76,769 $80,302
Avg need-based grant $58,675 $65,134 $70,764 $76,769 $80,302
% of need met 100% 100% 100% 100% 100%

Source: Williams CDS, H1/H2.

The average need-based grant for a first-year on aid has risen from $58,675 in 2021鈥22 to $80,302 in 2025鈥26, a 37 percent increase over four years. The structural piece is that this is a grant, not a package. Average package equals average grant under the all-grant policy. A family that qualifies for aid is not weighing $80K against $20K in loans they would take on; they are weighing $80K of grants against the sticker. For families with substantial demonstrated need, that is a category change in what Williams costs. For full-pay families, Williams remains a roughly $91,000 billed cost, about $96,000 all in once books, travel, and personal expenses are included (all of which the package covers for aid recipients).

One contextual point. Williams meets 100 percent of demonstrated need by its own methodology, including for international students. The college鈥檚 net price calculator (npc.collegeboard.org/app/williams) is the right starting point for any family doing serious financial planning, and the gap between what the calculator estimates and what the institutional methodology actually produces tends to be narrow. That methodology is generally more generous than the federal formula for middle-income families with non-trivial assets or self-employment income, and it is the one Williams uses.

7. Retention, graduation, and outcomes

Williams鈥 first-to-second-year retention has held within a narrow band of 96.9 to 97.4 percent across the five most recent cohorts.

Retention 2021鈥22 2022鈥23 2023鈥24 2024鈥25 2025鈥26
First-to-second-year retention 97.0% 97.2% 97.0% 96.9% 97.4%

Source: Williams CDS, B22. Each year reflects the prior fall鈥檚 entering cohort returning the following fall.

Six-year graduation rates carry the more important signal. For the Fall 2019 cohort (the most recent reported in the 2025鈥26 CDS), Williams graduated 95.2 percent of its entering class within six years. The breakdown by aid background is notable: 97.1 percent of Pell Grant recipients graduated, compared with 95.3 percent of subsidized Stafford recipients and 94.8 percent of students who received neither. Pell students graduated at a higher rate than students who received no need-based aid at all, an uncommon pattern. Most institutions show the opposite, with Pell graduation rates several points below the overall rate. Williams鈥 result is consistent with how it staffs and resources access programs, and it is a useful counterpoint to the assumption that aid-eligible students at highly selective colleges struggle to complete.

Williams鈥 student-faculty ratio is 7:1, and class-size data (CDS I3) shows that about three-quarters of sections enroll fewer than 20 students; in Fall 2025, 528 sections were offered with a mean of 15 students and a median of 13. The college鈥檚 outcomes profile (graduation, median first-job salary, fellowship and graduate-school placement) sits at the top of the national peer set, and the investment in the student-faculty model is what underwrites those numbers. The CDS confirms that, year after year, the model is being delivered.

8. What the data says, plainly

Five cycles of Williams CDS data, read together, sketch a college that has made one large institutional bet (the all-grant aid policy) and weathered the external shocks of the period without fundamentally altering its admissions posture. The admit rate has moved within a narrow single-digit band. ED admit rates have compressed but remain well above RD. The middle-50 test profile for submitters has crept upward as fewer students submit. Class size has grown modestly, financial aid generosity has grown significantly, and the racial composition of recent classes has shown only modest movement so far, with the second post-SFFA cohort the one to watch most closely.

A few practical takeaways for applicants and the counselors advising them. First, the admit rate has been single-digit every year in the window and will stay there; chasing yearly fractional movements in the published rate is not useful, and focusing on the strength of the file (rigor, GPA, recommendations, and the character signals Williams reads carefully) is. Second, the binding ED commitment is the most consequential strategic choice in the process. If Williams is the clear first choice, the ED math favors applying early; if it is not, the binding commitment is the real constraint, and waiting for RD is the right call. Third, the test-optional policy is truly optional, and the practical rule is simple: submit if scores sit at or above the 75th percentile, consider strongly in the 25th-to-75th range, and do not submit below. Fourth, the aid program changes what Williams costs in a category-altering way for aid-eligible families, including international students; the net price calculator is the right place to start any serious financial conversation.

Williams remains one of the most selective and best-resourced colleges in the country, and its CDS data shows an institution that knows what it is and how it wants to operate. The job of applicants and their families is to read that posture honestly, decide whether Williams is the right fit, and apply accordingly.