GMAT Superscore: GMAC’s Own Answers to What Schools Actually See
GMAT Superscore automatically assembles your highest Quant, Verbal and Data Insights section scores across your valid attempts into one Total, at no cost and with nothing to apply for. It...
GMAT Superscore automatically assembles your highest Quant, Verbal and Data Insights section scores across your valid attempts into one Total, at no cost and with nothing to apply for. It carries no percentile of its own. You cannot opt out of sending it, and every school on your list receives it. Build a single-sitting score you’d be content to be judged on alone, then let Superscore work in your favour.
A school can now see two GMAT scores from you.
The first is a 675 you earned in a single sitting. The second is a 705, built from your best Quant, your best Verbal and your best Data Insights, earned across three different Saturdays.
Which of those two tells an admissions reader more about you?
Don’t answer what you want to be true. Answer what a reader going through four thousand files would actually believe. Hold your answer — we come back to it.
GMAT Superscore is live. I put the questions students have been sending me straight to Debaion Roy, Associate Director of Marketing at GMAC for APAC, Middle East and Africa, on a live session, and what follows is his answers, GMAC’s own documentation and score report, and my reading of what all of it means for someone with a Round 1 deadline in September.
Six questions, in the order I would take them with a student sitting across from me.
Start with a free profile evaluation. We’ll look at where you stand and whether a retake even moves the needle.
What a GMAT Superscore actually is and which attempts feed it
Three sections, your best of each, assembled into one Total. No averaging, no estimation.
You have three sections: Quant, Verbal, Data Insights. GMAC looks across all your valid attempts, takes your highest Quant from any one of them, your highest Verbal from any one of them, your highest DI from any one of them, and assembles those three into one Total on the 205 to 805 scale.
Here is an example, so we work with real numbers.
| Quant | Verbal | DI | Total | |
|---|---|---|---|---|
| February | 79 | 85 | 82 | 645 |
| May | 85 | 81 | 82 | 655 |
| July | 83 | 83 | 85 | 675 |
| Superscore | 85 | 85 | 85 | 705 |
February gave you your best Verbal. May gave you your best Quant. July gave you your best DI. Your best single day was a 675. Your Superscore sits thirty points above the best day you ever had. Every one of those three section scores belongs to you, earned on a real test day under a real timer, with no averaging and no estimation anywhere in the arithmetic.
Seven mechanics to have straight.
There is nothing to apply for, nothing to buy, and no sections for you to hand-pick. GMAC’s system does the arithmetic and puts it on your report. The score report states that a Superscore is included with your score send at no additional cost.
GMAC generates a Superscore only where there is something to assemble. Sat once? No Superscore. Scored your highest Quant, Verbal and DI all in the same sitting? Also no Superscore, because assembling would just reproduce that attempt. Either way, your Official School Score Report states it plainly, so nothing reads as missing.
Every attempt and its scores remain part of your history, with the Superscore sitting alongside them as an additional data point. Each individual attempt stays separately sendable, exactly as before.
Whether taken at a test centre or online. A score from the old 200 to 800 exam contributes nothing, because the two scales were built on different psychometrics and cannot be combined. Expired scores are excluded as well, and current-edition scores only begin expiring in 2028.
The GMAT remains capped at eight lifetime attempts, so Superscore does not open an unlimited runway to optimise section by section.
If your Verbal was 86 in both February and May, GMAC uses the May one. Your Total stays where it was. The date shown against that section moves to May, and that date is visible to schools — we come to it in question five.
Open a Superscore report and each of the three sections shows its own percentile against its own score. The assembled Total shows none. That single detail settles the biggest worry students have about this launch, so it becomes question two.
GMAC’s own estimate of the effect is a 20 to 50 point gap between a first-attempt Total and a Superscore. That estimate averages across a very mixed population, so let me be more useful than the average. Fifty points does quite different work depending on where you are standing.
A student moving from 665 to 715 has changed the conversation, because a 715 reads differently at almost every programme they were considering. A student moving from 575 to 625 has improved a real amount and will be looking at broadly the same set of schools, so the honest reading there is that Superscore has not altered their strategy very much. I would also treat that range as a description of a population and not as a prediction about you.
Hearing “20 to 50 points” as twenty to fifty free points. That gap is the distance between your best single day and your best three half-days, and the only way to it is by sitting and paying for more attempts.
Treat 20 to 50 as a ceiling on the upside, then subtract what the attempts cost you in fees and in preparation weeks. If the number left over does not change which schools you can credibly apply to, then you are looking at an ordinary retake decision that has nothing to do with Superscore.
That is the machine. Now the worry it has generated.
Does this cause score inflation? What actually changes for the reader
The worry comes apart at the first step: percentiles have no route into an assembled Total.
Let me take this early, because everything downstream reads differently once it is settled.
The worry runs like this: more candidates now present higher numbers, so every number is worth slightly less, and today’s 705 becomes next year’s 675.
It is a reasonable line of thinking, and it comes apart at the first step. Open an actual Superscore report. Each of the three sections carries a percentile against its own score — a Verbal 75 might sit at the 18th percentile, a Quant 89 at the 97th. The Superscore Total carries no percentile at all.
GMAC computes percentiles from single-sitting scores, which Debaion confirmed on the call. So a Superscore has no route into the percentile tables and no mechanism for moving them. Your single-sitting score arrives with a rank attached to it. Your assembled Total arrives as a number on its own.
Sit with that for a second, because it is the most clarifying fact in this whole launch.
There is a separate point here that catches almost everybody. Your percentile drifts over the years regardless. GMAC refreshes those tables annually to reflect who is testing, and per its own documentation the shifts reflect population changes and broader trends in graduate management education. Your score stays fixed while its rank moves around it. That was true in 2019 and it is true this morning, and Superscore has nothing to do with it.
What does change is the size of the pool being seriously considered. Debaion’s phrasing was that Superscore widens the pipeline, so a school evaluating a hundred candidates might now be evaluating a hundred and forty. Then he said the part most people move past too quickly: your competition inside that pipeline goes up as well, so your profile matters more than it did before.
Reading “no percentile” as a downgrade, as though an assembled Total is a weaker currency than a real one.
Read it as the reason your single sitting still deserves your attention. It is the only one of your numbers that arrives with a rank, so it is the only one a reader can place against a cohort without doing any interpretive work. Build that number first. Let the Superscore sit on top of it.
When more people clear the number, the number does less of the filtering. What separates you moves further into your profile, your work and your essays. Which raises a fair objection about what the exam is now measuring at all.
Does this dilute the essence of the GMAT? A question worth voicing properly
Your single-sitting score is unedited on the report. What changes is who gets to weigh it, and against what.
Several candidates have written to me with this one, so let me voice it properly before answering it.
The GMAT is a timed, single-sitting, high-pressure assessment. Holding together under exactly those conditions is part of what it measures, and arguably part of what business school and the job afterwards will demand of you. Assembling a number from three calmer Saturdays could reasonably be said to remove some of that.
I put it to Debaion. His answer stayed with the single-sitting score, and structurally he is right. Your single-sitting performance is still on the report in full, unedited, so a school keeps every bit of its ability to judge how you held together across one continuous morning. Superscore has added a data point to a report that already carried everything it carried before.
Let me add the part that affects you directly. Your single sitting is intact and the section dates are visible. So a reader who cares mainly about single-sitting performance can read that and leave the assembled Total alone. Which means what Superscore really does is hand a choice to the school, and the school makes that choice without consulting you.
This is why Debaion’s own advice on the call was that your first priority should remain your single-attempt score. That is a notable sentence to hear from the organisation launching this, and it is the most useful thing anyone said in the whole conversation.
Planning a testing strategy whose only acceptable outcome requires a school to prefer your assembled Total. You have handed your plan to a decision you cannot see.
Build a single-sitting score you would be content to be judged on alone. Then let the Superscore be upside. Every good version of this plan works if the school ignores column two entirely.
So the exam still measures what it measured yesterday. Which leaves the question of why GMAC changed the reporting at all.
If a section is quietly costing you the points, our team can read your report with you before you book another attempt.
Why has GMAC introduced this? And what score anxiety actually costs
The problem GMAC is describing sits at the sending stage, not the scoring stage.
GMAC has given its own answer and I think it holds. Its documentation puts score anxiety among the most significant things keeping qualified candidates from sending their scores to programmes at all. The problem it is describing sits at the sending stage, well after the scoring is finished.
Look at the numbers underneath that, from GMAC’s own published class data.
| School | Median or average | Published range |
|---|---|---|
| Kellogg | 687 median | 515 to 775 |
| London Business School | 645 average | 555 to 805 |
| ISB | 675 median | 555 to 765 (middle 80%) |
Read the second column. Then read the third. ISB’s published range is the middle 80% of admitted students, which means 10% of the people ISB admitted scored below 555. Kellogg has admitted at 515. London Business School has admitted at 555.
Now picture the student holding a 655 who has decided a 655 rules them out of a school whose admitted range starts in the mid 500s. They hold the score back. They lose a cycle. The school never learns they existed. That is what score anxiety costs, and it costs the school something too.
There is a second reason worth understanding, and I raise it because reading an organisation’s incentives is a skill your MBA will assume you already have. A candidate who has one bad Saturday today has several low-friction exits available: the GRE, the CAT, an EA, or simply a cycle skipped. Superscore gives that candidate a reason to stay with the GMAT, because the good sections they already hold keep their value. It also makes a further attempt more appealing, which does a test provider no commercial harm at all.
You will have seen the version of this argument that says it is a business and it only wants more attempt fees. My own view is that an organisation can address a real candidate problem in a way that also serves its revenue, and both of those things sit together comfortably.
Letting the answer to “is this commercially motivated” decide what you do next. It is an interesting question about GMAC and a useless one about your September.
Ask the narrower question instead. Does the existence of Superscore change what you personally should do in the next sixty days? For most students reading this, it does not, and question five is where we work out whether you are the exception.
What does it mean for you? Going back to the answer you were holding
A school knows you sat more than once, and when. It does not know what you scored on the days you chose not to send.
Now go back to the answer you were holding. Here is what GMAC says a school actually receives. Three things.
Where a Superscore exists, showing the Quant, Verbal and Data Insights section scores used to build it, along with an explanation that it was assembled from the highest eligible sections of your valid current-edition attempts.
Test centre or online, attached to each of those section scores.
Every score you have sent that programme in the past five years.
Point b answers the question I have been asked most since this was announced, and it deserves a moment. You still choose which attempts to send. What travels regardless, where a Superscore exists, is the date and delivery method against each section that built it. So in our example, a school sees a Verbal dated February, a Quant dated May and a DI dated July. Three different dates against three sections tells a reader these came from more than one sitting. What the report withholds is your Total from each of those sittings.
So the school knows you sat more than once, and it knows when. It does not know what you scored on the days you chose not to send. Point c is the one students overlook entirely. If a 605 went to Booth in 2024, Booth still holds that 605.
Planning around concealment. Superscore hides the numbers from the attempts you didn’t send. It does not hide that those attempts happened.
Plan around what the whole report says together. A July 675 with a February Verbal and a May Quant reads as someone who sat three times and got stronger each time. That is a perfectly good story, and it only becomes a bad one if you were counting on the reader not knowing.
So what does this add up to for you? Three things.
Lifting your Quant by four points used to be worth nothing unless everything else held up on the same morning, and a strong Verbal in March could be undone entirely by a difficult DI in July. Both of those situations now work in your favour. This is the single most substantial thing in today’s launch.
A student who knows their Verbal is banked can spend eight weeks on DI alone instead of maintaining three sections at once. For a working professional with four hours a week, this may be the most valuable part of it, and it is where I would build a plan.
School positions differ, so improvement you assemble across attempts may be read closely at one programme on your list and set aside at another. You do not get to choose which. That is question six.
Before anyone books anything, though, there is one situation worth naming. Picture the student at 665, essays half written, three weeks from a Round 1 deadline, who reads about Superscore this morning and books a September date to push DI a little. Come October their file looks the same as it does today, and their application reads weaker, because those hours came out of their essays.
A retake triggered by a launch instead of by a diagnosis.
Write your reason down in one sentence. If the word “superscore” is what makes the sentence work, you have a reaction to this morning’s news and you should delete the booking. If the sentence reads “my DI is 76, my target schools are quant-heavy, and my ESR shows I lose four minutes on the first MSR set,” you have a diagnosis. Go and act on it with some confidence.
Which brings us to the part everyone wants a list for.
Which schools accept it? Correcting the question before answering it
There is no accepting to be done. Every school receives it. The real question is what each one does with it.
This is the question I am asked most, and it rests on an assumption worth correcting first.
There is no accepting to be done. Your Superscore is not something you submit for approval. It is calculated automatically, it cannot be switched off, and where one exists it travels with every score send you make at no additional cost. Every programme you apply to this cycle receives it. A programme that has publicly said it does not superscore still receives it, because the report is built by GMAC and not by the school.
Sit with that, because it flips the thing you were worried about. The question was never whether a school will accept your Superscore. The question is what each school does with a number already sitting on its desk. And on that, as of this morning, very few have said anything at all.
Let me show you the three kinds of page you will find, using the actual wording from three schools, because learning to read these pages is worth more to you than any list I could give you.
Yale SOM has said it plainly, and its language is about as clear as it gets:
Read what that actually commits to. Yale will look at it. It will look at it next to your individual sittings. And elsewhere on the same page it says it sees and considers every score reported to it while crediting the highest one you achieved.
Notice how well that matches the mechanics from question five. Yale is describing exactly what the report hands it: an assembled Total, your single sittings, and the dates. Its policy is a description of the document. This is the only major programme I have found using the word Superscore approvingly as of today.
Cornell Johnson is the example worth studying, because its page never says “Superscore” and still tells you something useful. Read both sentences:
The first sentence tells you what Cornell leads with, which is your strongest single sitting. Most people stop reading there.
The second sentence is the one that matters to you. “We may also review scores from other exams completed.” Johnson is telling you that your other attempts are in front of it and being looked at.
Work out what that means for a real file. Your February Verbal 85 sat inside a 645 you chose not to designate. Under a policy that reads only your designated attempt, that 85 is invisible. Under Johnson’s policy, it is on the desk and being read. Cornell has not committed to using the assembled Total. It has committed to seeing the sections that built it. For most students that distinction matters far less than it sounds, because the thing you were worried about was your good work going unseen.
Searching a school’s page for the word “superscore,” finding nothing, and concluding it will ignore your other attempts.
Search for what the page says about “other attempts” instead. “We may also review scores from other exams completed” and “we credit the highest score you achieved” are both commitments to look wider than one sitting, written before the word Superscore existed. Yale names the metric. Cornell describes the behaviour. Both are better news for you than silence.
Most programmes fall here, and there are two sub-cases.
Some have published nothing, because Superscore did not exist when their current guidance was written. Silence is a page that predates the launch.
Others carry a same-sitting requirement. Stanford’s test-scores page asks candidates to report scores from one examination and from the same testing session, and tells them not to average. Read carefully: that is a statement about which edition to read, not a statement about assembling sections across sittings.
Check the date on any rule before you interpret it. A rule written before this launch is a rule about something else that happens to sit near your question.
Why there is no list worth giving you yet
Debaion was direct about where the schools are on this. Some are still deciding, no reports carrying a Superscore have reached their desks, and GMAC is in the middle of explaining the metric to them. He called it year zero.
He is right, and here is the part to notice. Year zero is the cycle you are applying in. You will submit before most schools have finished working out what they think.
There is a fair amount of commentary circulating about which programmes will and will not use a Superscore, some of it from credible former admissions directors. None of it is school-published policy, and if this post named ten schools this morning and three published something different in September, it would have given you a reason to stop checking.
One reframe to carry while you wait. When a school eventually says it will not superscore, what it is usually describing is an operational constraint. These offices already reconcile GMAT, GRE, EA, two GMAT editions and several submission formats, and all of it has to stay comparable across four thousand files. A fresh composite metric arriving mid-cycle creates ambiguity inside a process built entirely on comparability. Workflow problems get solved on workflow timelines, which are slower than launch timelines. None of that is a statement about whether your improvement is real.
So what do you actually do?
Three things, and none of them involve waiting for a list.
Build your school list around fit and around your single-sitting score, exactly as you would have last month. Yale reading your Superscore does not make Yale a better fit for you than it was yesterday.
Assume, for planning, that the school reads your single sitting. Plan to a number you would be content to be judged on alone, and treat every Yale-style policy you find as upside.
Then read each target school’s own test-scores page in the week you submit, and read it the way we just read Cornell’s. Ask what the sentence is counting. In year zero, positions move faster than any third-party article gets updated, and that includes this one.
The GMAT measures today what it measured yesterday
What changed is how the measurement is displayed to a school. The work itself is unchanged.
It still comes down to reading one diagnostic honestly, naming the one weakness it shows you, and sitting one prepared attempt where all three sections get everything you have. If a second attempt is warranted after that, something in your method has to be different from the first, because an attempt with nothing changed behind it tends to reproduce the result you already had. Then the essays, which is where a widened pipeline gets decided.
And there is something in today worth carrying past this exam. Over the last few minutes you have watched a governing body change a reporting standard mid-cycle, seen schools respond to it in several different ways, and worked out what it means for you with incomplete information and real money at stake. That is close to the first case they will hand you in a business school classroom, where the job is to read the change, separate the mechanism from the commercial logic, work out what it means for your own position, and then move.
You did all of that this morning, before anybody taught you how.
Superscore has made the number a little more forgiving, and in doing so it has moved the weight of your application towards the parts that were always more interesting: what you have built, what you understand, and how clearly you can say so. That was already your strongest ground. You now have more room to stand on it.
The decision this article keeps pointing at needs one input you may not have yet: which section is actually costing you the points, and why. That’s a diagnostic question, not a scheduling one. Bring your ESR — we’ll tell you honestly whether a retake is worth your September.
Frequently Asked Questions: GMAT Superscore
Our experts have guided 30,000+ students through GMAT prep and MBA admissions. Get a personalised recommendation for your specific score and target schools.
Talk to a Crackverbal expert
Devmitra Sen is Head of Academics at Crackverbal and has trained over 4,000 students. Her scorers tell the story: GMAT 745, 725, 715, 705 alongside turnarounds like 575→715 and 375→675. She has produced multiple Q90 scores, including a perfect 100th percentile on GMAT Quant — a benchmark very few coaches can claim consistently. On Data Insights, her superpower is changing how students see, observe, and comprehend data: breaking it down, reasoning through it, and zeroing in on exactly what the question asks. The results follow: multiple 90+ percentile DI scores, including a 1st to 99th percentile turnaround in under two and a half months. She carries a quiet interest in the history of mathematical thought — particularly ideas rooted in India long before they were formalised elsewhere — a perspective that gives her an unusually grounded sense of why the subject matters.
TL;DR GMAT Multi-Source Reasoning (MSR) questions give you 2–3 tabs of data and ask you to cross-reference across all of...
TL;DR GMAT Graphics Interpretation (GI) questions appear in the Data Insights section and ask you to read a graph or...
TL;DR ISB’s Class of 2026 averaged 665 on the GMAT Focus Edition. There is no official cutoff. A score of...
Your 705+ score starts with
one expert call.
30-minute strategy session. We'll audit your prep, find your gaps, and build a roadmap to your target score.