Stanford Common Data Set: 2025-26 Numbers Decoded

The most useful admissions document Stanford publishes is a fill-in-the-blank questionnaire it did not write. Stanford stopped announcing acceptance rates on decision day years ago. It issues no press release in late March. The only place a real Stanford admit rate eventually surfaces is a standardized PDF filed with a survey consortium, months after the class is already enrolled. This page reads that file for you, line by line, and names the three numbers it will never contain.

Stanford's Common Data Set, in one screen

Stanford's most recent Common Data Set is the 2025–26 edition, covering Fall 2025 entry (the Class of 2029). It reports 60,646 applicants, 2,302 admitted (a 3.80% admit rate), and 1,839 enrolled. It is published by Stanford Institutional Research & Decision Support and is the university's only official source for its acceptance rate.

If you are applying in the 2026–27 cycle for Fall 2027 entry, this is the newest data that exists for you. Stanford released nothing public about the Class of 2030 when those decisions went out in spring 2026. That information arrives in the 2026–27 CDS, expected in early 2027, which will land after the Restrictive Early Action deadline you are currently staring at and probably after Regular Decision too.

The five numbers from the 2025–26 CDS worth memorizing

1. 60,646 first-year applications.

2. 2,302 admitted.

3. 3.80% admit rate.

4. 1,839 enrolled, the largest first-year class Stanford has reported.

5. 73% of enrolled first-years arrived with a 4.0 GPA.

Which edition is current right now (and when the next one drops)

As of August 2026, the newest Stanford CDS is 2025–26. Editions are named for the academic year they describe, so "2025–26" reports on students who applied in the 2024–25 cycle and enrolled in September 2025. The naming trips up almost everyone. When a blog post says "Stanford Common Data Set 2026-2027," it is either describing a file that does not exist yet or mislabeling the current one.

The pattern across recent years: Stanford posts the new edition in the winter or early spring following fall census. Expect the 2026–27 file (Class of 2030) roughly in the first months of 2027.

Where the file actually lives on Stanford's site

The archive is at Stanford Institutional Research & Decision Support's CDS page, with downloadable PDFs going back to 2008–09. That back-catalog is the part almost nobody uses, and it is where the interesting work is: pull five editions, put C1 side by side, and you can see what Stanford is actually doing rather than what any single year suggests.

What the Common Data Set is, and why Stanford's matters more than most

The Common Data Set is a standardized questionnaire completed voluntarily by colleges, developed as a collaboration between the higher education community and publishers including the College Board, Peterson's, and U.S. News & World Report. Every participating school answers the same questions in the same order, which makes institution-to-institution comparison possible in a way marketing pages never allow.

The Common Data Set Initiative sets the template. Schools fill it in. There is no auditor, no submission deadline, and no requirement to participate at all, which matters when you start comparing schools (more on that in the FAQ).

A shared questionnaire, not a Stanford document

This is the key mental shift. Stanford is not choosing what to tell you in the CDS. It is answering somebody else's questions. That is exactly why the file is more candid than any admissions webpage: the questions about waitlist size, transfer admits, and factor weighting were written by people who wanted comparability, not by a communications office.

Sections A through J, decoded

Section C is where you will spend 80% of your time. C1 is raw counts. C2 is the waitlist. C7 is the factor grid. C9–C11 are test scores and GPA. C21 is where early-round data would appear, if Stanford reported it.

Why Stanford stopped announcing acceptance rates

Stanford deliberately withholds decision-day admissions statistics as part of a stated effort to de-emphasize selectivity metrics. No admit rate on the day decisions go out, no running commentary during the cycle. The number surfaces only when the CDS is filed. For most universities the CDS is a supplement to a press release. For Stanford it is the only release, which is precisely why CDS literacy is worth more here than at almost any peer school.

Section C1: the acceptance rate and the trend behind it

Stanford's Class of 2029 admit rate was 3.80%: 2,302 offers from 60,646 applications, per the 2025–26 Common Data Set. The prior cycle (Class of 2028) was 3.61%, from 2,067 admits out of 57,326 applications. The rate rose slightly because Stanford expanded its admitted pool faster than applications grew.

Class of 2029 vs Class of 2028, side by side

The yield figures in that last row are arithmetic on the CDS counts, not a number Stanford advertises. Divide enrolled by admitted and you get roughly 80% for the Class of 2029 against roughly 82% the year before.

Applications up, admits finally up too

Across the five most recently reported cycles, Stanford's admit rate has sat in a band of roughly 3.61% to 4.34%, with the total number of offers pinned near 2,000 year after year. That flat admit count is the important structural fact. Application volume moved; Stanford's offer count barely did. Then the Class of 2029 broke the pattern with 2,302 offers and 1,839 enrolled, a genuine expansion rather than a rounding error.

Do not read that as "Stanford is getting easier." A move from 3.61% to 3.80% is noise at the individual applicant level. Read it as: the denominator grew, the numerator grew a little more, and the class got bigger.

Yield: why ~80% changes how you read every other number

When four out of five admitted students enroll, the admitted-student profile and the enrolled-student profile are nearly the same population. That matters for how you interpret C9–C11. At a school with 40% yield, "enrolled student SAT range" can differ meaningfully from "admitted student SAT range." At Stanford, they are close cousins. High yield also compresses the waitlist, which is the next section.

For context on how that 3.8% sits against comparable schools, our breakdown of Ivy League acceptance rates runs the same CDS-sourced math across all eight, and if you are wondering why Stanford is not on that list, we answered that separately.

Section C2: the waitlist, where the CDS gets brutally specific

Section C2 reports three separate waitlist numbers: how many applicants were offered a place, how many accepted one, and how many were ultimately admitted. For the Class of 2028, Stanford offered 483 waitlist spots, 414 students accepted, and 25 were admitted, a 6.04% waitlist admit rate. For the Class of 2029 that rate rose to roughly 9.95%.

Offered vs accepted a spot vs actually admitted

Those three numbers are not interchangeable, and almost every article that quotes a "waitlist acceptance rate" quietly picks whichever denominator flatters the story. The honest denominator is the number who accepted a place on the list, because those are the only people still in the running. Using 25 of 414 gives 6.04%. Using 25 of 483 gives 5.2%. Use the first one.

Why the waitlist rate nearly doubled

Waitlist movement is a yield artifact, not a signal about you. When yield slips (and Stanford's calculated yield fell about two points between those classes), the class comes up short and the waitlist opens. When yield holds, almost nobody moves. AdmissionSight's year-over-year read of Stanford's Section C2 puts the Class of 2029 waitlist admit rate near 9.95%, roughly double the prior year. That is not Stanford becoming generous. That is arithmetic on a class that filled slightly slower.

What that means if you land there in spring 2027

Plan for the 6% version, not the 10% version. Deposit somewhere you are genuinely happy to attend, then write one disciplined letter of continued interest and stop. Our letter of continued interest template covers what actually belongs in one (new academic results, a specific reason it is your first choice) and what gets ignored.

Section C7: Stanford's published rubric, read this before you write anything

Section C7 asks every college to rate 19 admission factors as Very Important, Important, Considered, or Not Considered. It is the closest thing to a published rubric any selective university offers. Stanford's grid places application essays in the top tier alongside academic rigor and GPA, and rates demonstrated interest and state residency as Not Considered.

The factor grid, as reported

Ratings can shift between editions, so open the current PDF on the IRDS archive and check C7 yourself before you build a strategy on it. It takes ninety seconds.

What "essays are Very Important" actually means for your drafts

Here is the benchmark that makes this useful. Across the roughly 104 schools indexed in GradGPT's Common Data Set explorer, rigor of secondary school record is rated Very Important far more often than the application essay is. Rigor is near-universal at this tier. The essay in the top tier is not.

So when Stanford puts essays in the same box as your transcript, that is a real signal about where the marginal decision gets made. Your transcript is largely fixed by August of senior year. Your essays are not. That asymmetry is the single most actionable thing in the entire document: the factor you can still change is the one Stanford rates highest among the changeable ones.

Practically, that means Stanford's short supplemental responses are not filler. They carry weight equal to the personal statement in a grid that does not distinguish between them. If you want a sense of how readers actually process those pages under time pressure, our piece on how admissions officers read application essays walks through it.

Factors rated Not Considered: stop optimizing for them

Four things in that bottom row are worth internalizing:

• Level of applicant's interest. Stanford does not track whether you opened the email, attended the info session, or visited campus. Demonstrated-interest strategy is wasted effort here. It is not wasted at plenty of other schools on your list, which is exactly why you should check each school's own C7.
• State residency. There is no California advantage encoded in the process, and the CDS reports no residency-based admit rates.
• Religious affiliation.
• Racial/ethnic status.

How to compare Stanford's grid against MIT's and Harvard's in ten minutes

This is the highest-value use of the CDS and almost nobody does it. Open MIT's institutional research CDS page and Harvard's Common Data Set in two tabs, jump to Section C7 in each, and build a four-column grid: factor, Stanford, MIT, Harvard. Then look only at the rows where the ratings disagree.

Those disagreement rows are your strategy. If one school rates interview "Important" and another rates it "Considered," you know where to spend your prep. If one rates level of interest "Considered" and Stanford does not, you know where to open the emails. Twenty minutes of this beats twenty hours of forum speculation, because you are reading what each school filed rather than what a stranger guessed.

Sections C9–C11: test scores and GPA, read the way an admissions officer would

Stanford's CDS reports middle-50% SAT and ACT ranges for enrolled first-years who submitted scores. Published figures circulating online range from 1510–1570 to 1520–1580 for the SAT and 34–35 for the ACT, depending on which CDS edition the source used. Under the test-optional years those ranges reflect submitters only, roughly half the class.

Why "of submitters" is doing heavy lifting

During Stanford's test-optional period, only about 50.3% of enrolled first-years submitted an SAT and about 19% submitted an ACT. Students with lower scores generally withheld them. That makes the published range a self-selected sample, skewed upward, and it is the reason you should treat those numbers as a description of who chose to submit rather than a threshold you must clear.

You will also see the ranges quoted inconsistently across the web (1510–1570 on one site, 1520–1580 on another). That is almost always an edition mismatch: one source is reading the 2024–25 CDS and the other the 2025–26 CDS. Check the file year before you trust the range.

The GPA bucket table and the tail below 4.0

Section C11 gives the distribution, not just an average. Roughly 73% of enrolled Stanford first-years reported a 4.0 GPA. Which means about a quarter did not, and they enrolled anyway.

That tail is the part worth staring at. It tells you that Stanford's process is not a GPA filter with essays attached. It also tells you that if you are sitting at a 3.85 with a genuinely brutal course load, you are inside the distribution, not outside it. The CDS does not tell you what those students had instead. The C7 grid does: rigor, essays, talent, character.

Test-required now: what the old ranges predict for Fall 2027

Stanford reinstated its standardized testing requirement, announced in June 2024 and effective for Fall 2026 entry. It is in force for the 2026–27 cycle you are in. So the honest question is: once everyone submits, do the ranges collapse?

Probably less than you would think. When submission rates were already running near 69–70% of the enrolled class in the later test-optional years, the non-submitting group was small enough that folding it back in shifts percentiles modestly rather than dramatically. Expect the 25th percentile to drift down somewhat and the 75th to stay roughly put. We work through the retake math and target ranges in detail in our Stanford SAT score requirements guide for Fall 2027.

Section D: transfer admissions, the page everyone skips

Section D of the Common Data Set covers transfer admission, and Stanford's numbers there are harsher than its first-year numbers. The 2025–26 CDS reports a Fall 2025 transfer admit rate of about 2.25%. Historical transfer rates have run near 1.8%, meaning transferring into Stanford is consistently more selective than applying as a first-year.

This is the section nearly every acceptance-rate article ignores, and it is the one that matters most to community college students, gap-year applicants, and anyone treating transfer as a backup route.

Why is it harder? Space. Stanford houses 100% of first-years in university housing and enrolls a class of roughly 1,800; transfer slots open only where attrition creates them, and attrition is minimal when first-year retention runs at 98.2%. There is no structural pipeline the way there is at large public systems with articulation agreements.

The practical read: transfer is a legitimate path for a specific kind of applicant with a distinctive record built in college, but it is not a softer second attempt. If you are choosing between an ambitious first-year application and a plan to transfer in later, the CDS says the first-year door is the wider one.

Sections G and H: what Stanford costs and what families actually pay

Stanford's published 2026-27 undergraduate tuition is $67,731, held flat from the prior year, with room and board at $22,944 (up 3.5%). The average university scholarship for students entering in fall 2025 was $69,415, and 86% of 2024-25 graduates left debt-free, according to Stanford's own announcements.

2026-27 sticker price, line by line

You will find total cost-of-attendance figures for Stanford online ranging from roughly $87,833 to $96,513. They conflict because they mix CDS editions, and because some include travel and personal allowances while others stop at tuition plus housing. Anchor on Stanford's own tuition and aid announcement rather than a third-party aggregator.

Average scholarship and the debt-free share

Section H reports institutional aid averages and the share of need met. The headline numbers are genuinely strong: an average university scholarship near $69,415 and the large majority of graduates finishing without loans. Those are population statistics.

Why Section H is not a substitute for the net price calculator

An average is not a quote. A family at $80,000 of income and a family at $220,000 both sit inside "average scholarship of $69,415" while receiving wildly different offers. Run the Stanford net price calculator with your actual numbers, then treat Section H as sanity-checking context.

And when real offers arrive next spring, compare them properly. Our guide to comparing financial aid packages covers the grant-versus-loan sleight of hand that makes two similar-looking letters differ by $20,000 a year.

The three things Stanford's Common Data Set will never tell you

Every article about the Stanford CDS lists what is in it. Here is what is structurally absent, which is arguably more useful.

1. No Restrictive Early Action vs Regular Decision breakdown. Section C21 exists for early-round reporting, and Stanford leaves the early admission data out. No REA application count, no REA admit count, no REA rate. Which also means no Regular Decision rate can be calculated, because you cannot subtract a number you do not have. Any site quoting a "Stanford early action acceptance rate" is publishing an estimate, a leak, or an invention.

2. No international or residency admit rates. The CDS reports enrolled demographics, not admit rates by citizenship. International applicants cannot compute their own odds from this file. State residency is separately rated Not Considered in C7, so the "California advantage" figures floating around are modeling, not data.

3. No admit rates by major or by school. Stanford admits to the university, not the department, and the CDS carries no per-major admit data. Section J tells you what percentage of degrees were conferred in each field, which is an outcome, not an admissions probability. Choosing a major to game the odds is optimizing against a number nobody has published.

A fourth gap worth knowing: the CDS uses the federal "two or more races" reporting rule, while some peer institutions publish demographics using a "check all that apply" method in their own press materials. Comparing Stanford's CDS race/ethnicity table to another school's press release can therefore mislead badly. Compare CDS to CDS or not at all.

What to use instead when the data doesn't exist

• For early-round strategy: reason from yield and policy, not from phantom rates. Stanford's REA is restrictive and non-binding, so it does not carry the yield-protection incentive a binding Early Decision program does.
• For international odds: use enrolled international share from Section B as a rough ceiling on class composition, and accept the uncertainty.
• For by-major questions: look at Section J degree distribution and department size for capacity signals, and stop there.

Turn the CDS into a plan: a 7-line self-audit against Stanford's own grid

The grid only earns its keep if you score yourself against it. Copy these seven lines, rate each 1 to 5, and be unkind.

Read your two lowest scores. Those are your November project. In practice, for most students applying this fall, the two lowest are lines 3 and 7, and they are the same problem: the essays are not carrying the character evidence.

What fixes what, before November 1: rigor and GPA are locked. Recommendations are semi-locked (you can still brief your recommenders with specifics this month). Essays, talent framing, and character evidence are fully in play for the next ten weeks. Spend your time proportionally.

Then run the same read on the other 8 to 12 schools on your list. Pull each school's C7, note where their Very Important column differs from Stanford's, and adjust. A school that rates interview "Important" deserves prep time Stanford does not. A school that counts demonstrated interest deserves an opened email.

When the numbers say reach: at 3.8%, Stanford is a reach for every applicant alive, including the 4.0 with national awards. Build a list where the reaches are genuinely reaches and the likelies are genuinely likely, and treat any admit rate under 10% as a lottery you have earned a ticket to rather than a target you can hit reliably.

Run the audit with something that has read the file

This is exactly the work Unive was built for. Our Profile Evaluator scores your academics, activities, and essays against the factors selective schools actually publish, then the Application Roadmap turns the gaps into dated tasks against your real deadlines. The Essay Grader gives line-level feedback on the one Very Important factor you can still change before November 1, and College Matching builds a list where the reach/target/likely split is honest rather than aspirational.

Unive is built by Yale graduates, runs on a paid subscription with a 7-day money-back guarantee, and costs a fraction of a private consultant. If you want to see what the essay side looks like first, our college essay grader walkthrough shows the kind of feedback it returns.

Stanford deadlines and testing rules for the cycle you're in now

For Fall 2027 entry, Stanford's Restrictive Early Action deadline is November 1, 2026, and the Regular Decision deadline is January 5, 2027. Standardized testing is required for this cycle following the reinstatement announced in June 2024 and effective from Fall 2026 entry. Stanford superscores across test sittings.

Restrictive Early Action: November 1, 2026

REA is non-binding. You may still apply Regular Decision elsewhere, but the restriction limits other private early applications, so read Stanford's own policy page rather than a forum summary before you commit. Decisions typically release in mid-December.

Regular Decision: January 5, 2027

Decisions release by early April 2027. Waitlist movement, if any, resolves through late spring.

Testing required, and how superscoring works

You must submit SAT or ACT scores. Stanford considers your highest section scores across sittings, so a strong second attempt on one section does not get discarded because the composite was lower that day. Practically: if you have a splashy Reading score in March and a splashy Math score in June, both count.

One temporal warning. A meaningful share of pages ranking for Stanford CDS queries right now are still framed to the 2025–26 application cycle, and a few still describe Stanford as test-optional. Check the publication date on anything you read about deadlines, including this page.

FAQ

Is there a Stanford Common Data Set for 2026-2027 yet?

No. As of August 2026, the newest edition is the 2025–26 Common Data Set, covering Fall 2025 entry and the Class of 2029. The 2026–27 edition, which will report Class of 2030 admissions data, is expected in early 2027. Stanford publishes no admissions statistics between decision day and the CDS release, so nothing about the Class of 2030 is officially available yet.

Does the CDS show Stanford's early action acceptance rate?

No. Stanford does not report Restrictive Early Action application or admit counts in Section C21 of its Common Data Set. Because the early-round numbers are absent, a Regular Decision rate cannot be calculated either. Any published Stanford REA acceptance rate is an estimate rather than institutional data, and should be treated with skepticism no matter how precisely it is stated.

What GPA and SAT do I need based on the CDS?

The CDS reports distributions, not requirements. About 73% of enrolled Stanford first-years had a 4.0 GPA, meaning roughly a quarter did not. Published middle-50% SAT ranges run between 1510–1570 and 1520–1580 depending on the edition, and those reflect submitters only. Stanford publishes no minimum GPA or score, and C7 rates rigor and essays alongside grades.

Is the CDS more reliable than U.S. News or Niche?

Yes, for admissions figures. The CDS is filed by the university's own institutional research office in a standardized format, while rankings sites and review platforms reprocess, estimate, or crowdsource. Rankings publications helped design the CDS and draw on it. The tradeoff is timing: the CDS arrives months after decisions, so it is more accurate but less current than a rumor.

Do all colleges publish a Common Data Set?

No. Participation is voluntary. Most selective US universities publish one, but some decline entirely, some omit individual sections, and some post months later than their peers. That is why cross-school comparisons can silently break: a missing C2 or C7 is not a zero, it is an absence. Always confirm you are comparing the same edition year across schools before drawing conclusions.