Is the GRE Required for a Master's in AI in 2026?

For most master's programs in AI, machine learning, and computer science in 2026, the GRE is either optional or not required at all. That has been the direction of travel since around 2020, when a wave of departments dropped the requirement and many never brought it back. There are still exceptions, a handful of programs (some in the US, more common in a few Asian systems) that require or strongly recommend a score, so the honest answer is: usually no, sometimes yes, and you have to check each program individually. This post is informational, not legal, immigration, tax, or financial advice, and admissions rules change, so verify on the program's own page before you book a test.

What "required" vs "optional" vs "waived" vs "test-blind" actually means

The confusion is that four different policies all get lumped under "do I need the GRE." They are not the same thing, and the difference decides whether you spend roughly $220 and two months of prep or skip it entirely.

• Required: no score, no complete application. The system will not let you submit, or the committee rejects incomplete files. Rare now for AI master's, but it exists.
• Recommended: technically optional, but the program signals that strong applicants tend to submit. Read this as "optional with a nudge," especially if the rest of your file is average.
• Optional (test-optional): you choose. If you submit, they read it; if you do not, they will not hold it against you (that is the stated policy, anyway). This is the most common stance for AI and CS master's in 2026.
• Waived: the requirement is dropped for a specific cohort or window, sometimes case by case (for example, applicants from a partner university, or during a particular admissions cycle). A waiver is narrower than a blanket policy.
• Test-blind: they will not look at a score even if you send one. Less common than test-optional, but it removes the decision entirely.

The GRE itself is a general graduate-admissions test measuring verbal, quantitative, and analytical writing, and scores stay valid for five years, per the official ETS GRE page. So a score you took for a different plan two years ago may still count.

The regional split: where the GRE still shows up

Where you apply matters more than what you study. The pattern, roughly, as of 2026:

• United States: most AI and CS master's went test-optional and largely stayed there. A few programs still require or recommend it, and some funded or research-track master's lean on it more than coursework master's. Assume optional, confirm per program.
• Canada: the GRE was never standard for most Canadian CS master's and is rarely required now. Some research-based programs list it as optional or recommended.
• UK, Ireland, and most of Europe: the GRE is generally not part of the process. Admission usually rests on your bachelor's degree classification, transcript, and sometimes an interview. Continental European AI master's almost never ask for it.
• Parts of Asia: more mixed. Some programs in a few systems still expect a GRE, so if you are applying there, check early.

If you are weighing a US master's against a European or Canadian one, the testing burden is one more input in that decision. Our related read on whether a US CS master's is still worth it after the H-1B changes covers the bigger trade-off, and you can compare all 21 countries on cost, visas, and pay in the full guide.

When submitting a strong score still helps

Optional does not mean useless. There are specific cases where a good GRE (broadly a strong quantitative score for AI, say the mid-160s and up, though targets vary by program) can genuinely move your application:

• Your GPA is on the low side. A high quant score is the cleanest way to signal you can handle graduate math after a weak transcript.
• Your degree is from a school the committee does not know. A standardized number gives them a common yardstick.
• You are a career switcher coming from a non-CS or non-quant background and want to prove numerical readiness.
• You want funding. Some assistantships, fellowships, and scholarship committees still weigh test scores even when admission does not. If money is on the line, a score can pay for itself.
• The program says "recommended." Treat that as a soft yes if your file is otherwise unremarkable.

Flip side: if you have a strong GPA from a known program, relevant projects, and good letters, a GRE score adds little and a mediocre one can quietly hurt you at a test-optional program. Do not submit a weak score just to fill a field.

How to check each program and decide this week

Do not rely on a Reddit thread or a two-year-old blog. Confirm the current policy at the source, then decide. Here is the order that saves the most wasted effort:

1. Open the program's own admissions page (not the university homepage, the specific MS in AI/ML/CS "requirements" or "how to apply" page). Search that page for "GRE."
2. Classify what you find into one of the five buckets above: required, recommended, optional, waived, or test-blind. If the page is silent, email the graduate admissions office and ask directly; silence is not the same as "not required."
3. Check for a waiver path. If it is required or recommended, look for waiver criteria (GPA threshold, partner school, work experience). You may already qualify.
4. List your target programs in a spreadsheet with a policy column. If even one program you seriously want requires it, you probably take the test. If none do, you likely skip it.
5. Mind the deadline math. Scores take about 8 to 10 days to reach you and programs, and you can sit the test roughly every 21 days, so if you decide to test, book it at least six to eight weeks before your earliest deadline. Details are on the official ETS GRE scheduling page.

English tests are a separate requirement from the GRE and are far more commonly mandatory, so do not confuse the two. See our guide to IELTS and TOEFL requirements for studying AI abroad for that side.

The honest takeaway

For the median AI master's applicant in 2026, the GRE is optional and skippable, and your time is better spent on projects, your statement of purpose, and letters. Take the test if a program you want requires or recommends it, if you need to offset a weak GPA or an unknown school, or if you are chasing funding that weighs scores. Skip it if your file already tells a strong quantitative story and every target program is test-optional or test-blind. The tie-breaker is always the program's current official page, not last year's advice. Building a country and program shortlist first tells you fast whether any of your targets even want a score, which is exactly what the AI Relocation Guide is for.

Rule of thumb: check each program's own page, and only take the GRE if a target requires it, recommends it, or if a strong score fixes a real weakness in your file.