CS PhD Acceptance Rates: Stanford, MIT and CMU
Official CS PhD acceptance rates: Stanford CS admitted 107 of 2,645 (4.0%) for 2025-26; MIT EECS 169 of 4,057 (2026 on its dashboard). CMU: none.
CS PhD Acceptance Rates: Stanford, MIT and CMU
Stanford's Computer Science PhD admitted 107 of 2,645 applicants (4.0%) for 2025-26 entry. MIT EECS admitted 169 of 4,057 applicants in the year MIT's dashboard labels 2026 (4.2%, our arithmetic). Carnegie Mellon publishes no CS PhD acceptance rate. Those are the only official numbers we could find for the three programs.
Percentages for all three schools circulate on guide sites and forums without a source. This page uses only the universities' own dashboards, says which ratios we computed, and explains one trap: MIT's own dashboard shows EECS at 10.3%, a figure that blends in an internal MEng row. If you are deciding where to apply, the rate matters less than what it implies for your PhD statement of purpose: at about 4%, our view is that the fit paragraph does the most work, for reasons set out below.
Every dashboard and page below was read on October 29, 2026. Stanford and MIT update these dashboards once a year, so check the year label before you quote a number.
On this page: the numbers side by side · Stanford CS, five years · MIT EECS and the 10.3% trap · CMU · the national Taulbee figures · what 4% should change · questions
The official numbers, side by side
| Program | Year, as the dashboard labels it | Applied | Admitted | Admit rate |
|---|---|---|---|---|
| Stanford CS PhD | 2025-26 | 2,645 | 107 | 4.0% (Stanford's figure) |
| MIT EECS (department row) | 2026 | 4,057 | 169 | 4.2% (our arithmetic) |
| MIT EECS (department row) | 2025 | 4,478 | 190 | 4.2% (our arithmetic) |
| CMU Computer Science Department PhD | n/a | not published | not published | not published |
Stanford labels each year by the academic year the applicant meant to enroll, so its latest column is 2025-26. MIT's dashboard labels years with a single number and does not define it; its latest column is 2026. We show MIT's 2025 column too because, as explained in the MIT section, it appears to line up with Stanford's 2025-26. That pairing is our reading, not MIT's. On it, both programs sit at about 4%.
The counts are not quite the same kind of number at each school:
- Stanford counts people. Its doctoral admissions page says "Prospective students may only apply to a single doctoral program at a time" (the 14 Biosciences programs are the exception). So Stanford's 2,645 applications are close to a count of people.
- MIT counts applications to EECS graduate study. EECS says "there is only one admissions process. When we admit an applicant, it is into the PhD program" (EECS admissions FAQ). The department row, without the separate MEng row, is therefore the PhD-track intake.
- Neither is a national rate. For that, see the CRA section below, and note that it counts applications, not people.
Stanford CS PhD acceptance rate, five years
Stanford's Institutional Research & Decision Support (IR&DS) dashboard lets you pick a single doctoral program. For "School of Engineering — Computer Science" it shows 2,645 applied, 107 admitted, 69 matriculated for 2025-26, and these admit rates:
| Entry year | CS PhD | Electrical Engineering PhD | All School of Engineering doctoral programs |
|---|---|---|---|
| 2021-22 | 6.4% | 12.5% | 9.1% |
| 2022-23 | 6.9% | 10.1% | 9.6% |
| 2023-24 | 6.4% | 9.5% | 9.2% |
| 2024-25 | 4.6% | 8.9% | 7.8% |
| 2025-26 | 4.0% | 7.0% | 5.9% |
Electrical Engineering's 2025-26 counts are 1,008 applied, 71 admitted and 53 matriculated.
How Stanford counts, in its own words:
"Note that any year referenced in this dashboard refers to the academic year in which the applicant was intending to enroll. For example, an application submitted in September 2018 for the 2019-20 academic year would be counted under 2019-20. These data are limited to new, external applicants only."
Two consequences. Current Stanford students who move into the PhD without a new application are not in these numbers. And the page notes that because the dashboards are updated each autumn, "the data for the most recent year will not include applicants or admits for winter or spring." The 2025-26 row could still move slightly.
Our arithmetic: 69 of 107 admitted students enrolled in 2025-26, or 64.5%.
MIT EECS PhD acceptance rate, and the 10.3% trap
MIT Institutional Research publishes a Graduate Education Statistics dashboard. With the department set to "Electrical Eng & Computer Sci", its Graduate Admissions table shows two rows for EECS:
| Year on dashboard | EECS applied | EECS admitted | Admit rate (our arithmetic) |
|---|---|---|---|
| 2026 | 4,057 | 169 | 4.2% |
| 2025 | 4,478 | 190 | 4.2% |
| 2024 | 4,529 | 209 | 4.6% |
| 2023 | 4,033 | 231 | 5.7% |
| 2022 | 3,560 | 235 | 6.6% |
| 2021 | 3,709 | 221 | 6.0% |
| 2020 | 3,749 | 242 | 6.5% |
| 2019 | 3,634 | 234 | 6.4% |
| 2018 | 3,467 | 215 | 6.2% |
| 2017 | 2,966 | 230 | 7.8% |
The second row is labelled "EECS (M.Eng Program)". In 2026 it reads 278 applied, 278 admitted and 240 enrolling. That is MIT's Master of Engineering program, which EECS describes as "Open to EECS undergraduates in 6-1, 6-2, 6-3, 6-4, 6-5, 6-7, and 6-14 programs" (EECS MEng page). It is not the route an outside applicant uses.
The trap. The dashboard's chart labels EECS's 2026 "Admitted Rate" as 10.3%. That figure matches the two rows added together: 447 admitted (169 + 278) out of 4,335 applications (4,057 + 278), which is 10.3% (our arithmetic). The dashboard's 78.7% "Yield Rate" matches the same blend: 352 enrolling (112 + 240) out of 447. Quote the department row, not the chart label, if you want the rate an outside PhD applicant faces.
The year label. The dashboard does not define its years. Its "All MIT" row for 2025 (39,982 applied, 3,752 admitted, 2,270 enrolling) is within a few of MIT Facts' "Selected Fall 2025 Graduate Admissions Statistics" (39,985 applications, 3,754 offers, 2,275 registered). So we read "2025" as fall 2025 entry and "2026" as fall 2026 entry. That is our inference; MIT does not say so on the dashboard, so this page labels MIT's years as the dashboard does. MIT Facts' institution-wide rate is 9%; that is all graduate programs, not EECS.
The EECS FAQ adds two lines that matter for applicants: "Admission is extremely competitive," and, on extra material, "Because of the large volume of applications submitted to EECS, we must discourage you from sending in copies of awards, journal articles, and other extraneous documents."
CMU CS PhD: why there is no official number
We looked for a Carnegie Mellon CS PhD acceptance rate on October 29, 2026 and did not find one. What we checked:
- The CSD doctoral admissions page: requirements, deadlines and funding, but no applicant or admit counts.
- The SCS graduate admissions FAQ: no counts.
- The SCS master's programs comparison: selectivity for master's programs only.
- CMU's 2026 Quick Facts: cohort admissions printed beside high-school GPA and SAT figures, and no doctoral admissions counts.
So any CMU PhD percentage you see elsewhere is not from CMU. We do not print one. CMU does publish master's selectivity; our CMU MS CS acceptance rate page covers those numbers and what each master's statement asks for.
What CMU's Computer Science Department does publish is useful for the decision: "A master's degree is not required to apply for the Ph.D. in Computer Science," and the department "offers all admitted Ph.D. students comprehensive financial support (stipend, tuition, university student fees, university student health insurance…) while they are in good academic standing." Its statement of purpose is "a concise one- or two-page essay describing your primary research interests, related experiences and objective in pursuing a Ph.D. in computer science." Fall 2027 deadlines on the same page: early application November 18, 2026 and final December 9, 2026, both 3 p.m. EST.
The national number: CRA Taulbee Survey
The Computing Research Association's Taulbee Survey is the only national source we found that reports CS PhD applications and acceptances.
- 2024 report (Table D14, "to begin in 2024-2025 Academic Year", 101 US CS departments): 50,667 applications and 5,847 acceptances. CRA does not print a rate. Our arithmetic: 5,847 ÷ 50,667 = 11.5% of applications. International: 4,242 ÷ 41,010 = 10.3%. Domestic: 1,605 ÷ 9,657 = 16.6%.
- 2025 report (doctoral section, 2025-26 entry): 47,267 applications reported by 125 units and 5,955 acceptances reported by 141 units, including computer engineering and information programs and Canadian units. Because the two counts come from different sets of respondents, we do not divide them.
Read the 11.5% as a share of applications. Each department reports the applications it received, so one person who applied to eight departments appears up to eight times. It is not your personal chance, and it says nothing about any one program. The 4% figures above are program-level; the national share is not.
What a 4% rate should change about your application
Our view: a 4% admit rate is a reason to make your research match checkable, not a reason to send more. Stanford admitted 107 people into one CS PhD program, and MIT EECS admitted 169. Those admits are spread across many research groups, and MIT says "Research assistants are chosen by faculty who are seeking students who share their research interests." At that scale, the statement's job is to let a specific faculty reader see where you fit in under a minute.
Once you have offers, PhD advisor red flags you can check covers the records to read before you accept.
The basis is the programs' own wording, not a model of how committees vote. MIT asks for two essays, one on "Research Activities and Goals" and one on your "Academic and Intellectual Journey" (our MIT EECS essays guide covers how to split them), and discourages extra documents. CMU asks for research interests in one or two pages.
The countercase: an admit rate measures the pool, not you, and not the training. A program with a 12% rate can be a better match for your question than one at 4%. If your subfield has few faculty at a school, its overall rate tells you little about your chances there. And none of these sources measures fit, advising or outcomes.
A procedure you can run with this page:
- Treat the rate as fixed and the fit as the variable. You cannot change 4%. You can change whether a reader in your subfield recognizes your question.
- Name the research group and the reason. One or two faculty, each with a sentence on their current work and how your project connects. Our CS statement of purpose guide shows what CS departments ask for.
- Spread the list by evidence, not by guesses. Use published program data where it exists and say "unknown" where it does not. How many PhD programs to apply to covers list size, fees and tailoring time.
- Do not pad the file. MIT asks you not to send extraneous documents. Put links where the form allows them.
- Check the deadlines. PhD application deadlines for fall 2027 lists the closing dates.
If you are weighing a master's first, read PhD straight from undergrad: MIT, Stanford and CMU all take bachelor's applicants into the PhD. If you are already in a US master's, the MS to PhD internal transfer list shows which CS departments have a route.
How this page was built
- We opened Stanford's IR&DS admissions dashboard and MIT IR's graduate dashboard on October 29, 2026, selected the program or department, and recorded every count shown here from the screen. We kept screenshots.
- Stanford publishes the CS admit rates in the table. Every MIT rate is our division of MIT's counts.
- We searched CMU's department, school and institutional pages for a PhD figure and found none.
- We did not use forum reports, consultancy estimates or self-reported admit lists. No figure on this page is an estimate.
Questions about CS PhD acceptance rates
Is MIT EECS harder to get into than Stanford CS?
On the published counts, they are about the same. In MIT's 2024 column, EECS admitted 4.6% (our arithmetic); in Stanford's 2024-25 column, CS admitted 4.6%. In MIT's 2025 column and Stanford's 2025-26 column, the figures are 4.2% and 4.0%. Pairing those columns is our reading of MIT's year label. The counting also differs (MIT counts applications to EECS, Stanford counts single-program applicants), so treat them as similar, not ranked.
What is the CMU CS PhD acceptance rate?
CMU does not publish one. We checked the CSD admissions page, the SCS graduate FAQ, the SCS master's comparison and CMU's 2026 Quick Facts on October 29, 2026.
Does a master's degree raise my chances?
None of these sources measures that. Stanford, MIT and CMU all accept bachelor's applicants to the PhD. MIT says that if you already hold a master's from another school, "you can begin working directly toward the PhD qualifiers."
Are international applicants admitted at lower rates?
Stanford and MIT do not publish admit rates by citizenship. Nationally, the CRA's 2024 report (2024-25 entry) counts 4,242 acceptances from 41,010 international applications and 1,605 from 9,657 domestic ones across 101 US CS departments: 10.3% against 16.6% (our arithmetic). These are shares of applications, not of people.
What GPA does Stanford CS expect?
Stanford's CS PhD FAQ says: "Typically GPA's are at least 3.6 (on a 4.0 scale) for PhD applicants." It also says GRE scores "are not required and not considered from PhD applicants." Admission at all three departments comes with funding; see how Stanford, MIT and CMU fund admitted PhDs, including international students.
Get a free review of your PhD statement of purpose
Get a free Quick Review of your PhD statement of purpose against GradPilot's adviser-fit rubric. It reads whether your statement names a focused research direction, ties it to a person, lab or area in the program with checkable details, and shows what your own work would contribute. For the whole statement, use the general PhD statement of purpose rubric; applying to master's programs instead, use the MS SOP rubric for computer science, AI and data. The PhD SOP review page explains how it works.
More PhD guides: the PhD essays hub collects every statement, list-building and offer guide in one place.
Sources
Retrieved October 29, 2026.
- Stanford IR&DS, Doctoral Admissions (methodology text) and its embedded admissions dashboard, programs "School of Engineering — Computer Science" and "School of Engineering — Electrical Engineering" (2021-22 to 2025-26 entry): counts, admit rates, definitions.
- MIT Institutional Research, Graduate Education Statistics dashboard, department "Electrical Eng & Computer Sci", school "School of Engineering" and "All MIT" (years labelled 2017 to 2026): EECS and MEng counts, chart labels.
- MIT Facts, Graduate Admissions (fall 2025): institution-wide counts used to read the dashboard's year label.
- MIT EECS, MEng program (no year stated): who the MEng is open to.
- MIT EECS, Graduate Admissions FAQs (no year stated): single admissions process, competitiveness, extra documents, essays, research assistants.
- Carnegie Mellon Computer Science Department, Doctoral Admissions (fall 2027 deadlines): master's not required, funding, statement of purpose.
- Carnegie Mellon SCS, Graduate Admissions FAQ and Master's Programs Comparison; CMU IR&A, Quick Facts 2026: checked for PhD admissions figures; none found.
- Stanford Computer Science, PhD Admissions FAQ (no year stated): GPA and GRE lines.
- Computing Research Association, 2024 Taulbee Survey, Table D14 (2024-25 entry), and 2025 Taulbee Survey, doctoral programs, Tables D14.a and D14.b (2025-26 entry).
Get feedback on your statement of purpose
Paste your draft. The free Quick Review shows what to fix against an admissions rubric before you submit.