GMAT Master Guide 2026

What the GMAT measures

The GMAT is a graduate management admissions exam used by business schools as one part of an application. It tests reasoning with quantitative information, written arguments and data presented in several forms. It does not test business knowledge or ask whether you have already studied accounting, finance or management. A strong result can support an application, but schools consider it alongside academic records, work experience, essays, recommendations and other requirements.

There is no single score that counts as a pass. A useful target is a score competitive for the programs on your list, given their published class profiles and your other application evidence. A score average is descriptive, not a cutoff that guarantees admission. Schools may also treat section strengths, work history and academic preparation differently.

Current format and timing

You choose the order of the three sections. Each section has its own 45-minute clock. An optional 10-minute break may be taken after the first or second section, and it does not count toward the 135 minutes of testing. Choosing an order can help manage energy, but it does not change section scoring or reduce the questions in another section.

Quantitative Reasoning contains 21 Problem Solving questions. It focuses on arithmetic and elementary algebra. You cannot use a calculator in Quant, so learn to estimate, simplify fractions and keep arithmetic organized. Verbal Reasoning has 23 questions in Reading Comprehension and Critical Reasoning. Data Insights has 20 questions across Data Sufficiency, Multi-Source Reasoning, Table Analysis, Graphics Interpretation and Two-Part Analysis. The on-screen calculator is available in Data Insights.

Adaptive questions and pacing

The GMAT adapts question difficulty as you work through a section. Your score is not a simple count of correct answers. The scoring model considers the difficulty of questions answered correctly and incorrectly, the pattern of answers and unanswered questions. Because of that, a practice-set percentage cannot be converted directly into a predicted total score.

Pacing matters because an unanswered item can hurt your score. Do not spend several minutes trying to rescue one difficult question while leaving later questions blank. A practical habit is to notice the time at regular checkpoints, make a reasoned selection when stuck and keep moving. The best checkpoint schedule depends on the section and your pace; practice with the full 45-minute clock rather than relying on an assumed seconds-per-question rule.

At the end of each section, Question Review and Edit lets you revisit bookmarked answers if time remains. You can change up to three answers in that section. You cannot use this feature to return to a section after beginning the next one. Treat it as a final check for a specific uncertainty, not as a second attempt to solve the entire section.

How GMAT scores work

The current total score ranges from 205 to 805 in increments of 10, with scores ending in 5. Each of the three section scores ranges from 60 to 90 in one-point increments. Sections carry equal weight in the total score. A total score is therefore useful for comparison, but section scores show whether your performance is balanced or whether one area stands out.

The score is scaled and adaptive. It does not mean that a 645 represents a particular percentage correct. Two candidates with the same raw number of correct answers can receive different scores if the difficulty and answer pattern differ. Do not use a percent-correct calculation from a small practice set to estimate your official total.

Percentiles compare a score with a reference group and can change when the score distribution is updated. Use the current official percentile table when interpreting a score. A percentile is not a probability of admission. For instance, a high percentile does not promise an offer, and a lower percentile does not determine an outcome without considering program context and the rest of the application.

GMAT scores are valid for five years. Official results usually arrive within three to five days, although processing can take as long as 20 days. The report contains section and total scores and additional feedback about performance. Keep enough time before application deadlines for a result to arrive and for any planned retake.

Superscore and sending results

GMAT Superscore combines the highest section scores from eligible valid attempts. It launched on August 12, 2026. The score report shows the superscore alongside a valid single-attempt score. The superscore has no separate percentile, so do not compare its percentile with the percentile for an individual sitting. Schools decide whether and how to use superscores in admissions.

You select programs to receive official scores through the score-sending process. Five free score sends are available within 48 hours after the official result is available. Additional recipients have a fee. Review each school's score-reporting instructions and deadlines before selecting recipients. A school may ask for all scores or may have its own policy about superscoring, so understand those requirements before you plan an attempt strategy.

Registration, appointments and costs

Registration starts with an mba.com account and a complete profile. You choose an in-person test center or online delivery, select an appointment, and pay during booking. The exam is offered throughout the year, but that does not guarantee a specific date or location will have a convenient opening. Schedule around application deadlines, preparation time and any planned retake interval.

The official price depends on country and delivery method. Since the payment page calculates price after location and delivery selections, use the amount displayed for your actual booking rather than an old fee quoted on another page. Budget separately for possible rescheduling, a retake, additional score reports and any optional preparation materials. A retake is another paid appointment.

Appointments must be at least 16 calendar days apart. Candidates may take the GMAT up to five times in a rolling 12-month period, combining online and test-center attempts. There is no lifetime attempt limit under the current policy. A perfect score of 800 on the former scale or 805 on the current scale creates a five-year wait before another attempt.

A useful way to study

Start with a timed diagnostic across all three sections. Record more than the score: note question type, time spent, whether you guessed, and why the chosen answer seemed right. Then sort errors into knowledge gaps, reasoning mistakes, careless execution, and pacing. A wrong answer caused by confusing a ratio with a difference needs a different lesson from one caused by running out of time.

Build a weekly loop. Learn or refresh one skill, solve a short untimed set, explain each answer, then return to a timed set several days later. Use new questions to test transfer. Repeating the same remembered question may measure recall rather than skill. In Quant, write clean equations and check units. In Verbal, identify the claim and evidence before evaluating choices. In Data Insights, read labels, units and conditions before calculating.

A full-length practice session helps test stamina and section order. Review it slowly afterward. For every miss, identify the exact point where the reasoning changed course. For a correct answer reached by luck, write down the uncertainty too. Track a small number of recurring error patterns so practice changes behavior rather than becoming a sequence of score reports.

Suppose you study ten hours each week for eight weeks. That is 80 scheduled hours, but the useful total depends on how much time includes deliberate solving and review. If four of those hours each week are spent replaying explanations passively, you may have fewer than 50 hours of active practice. A better log records focused blocks and the question type reviewed, not just time with a book open.

Original worked example

A supplier charges a fixed delivery fee of $18 plus $4 for each item. A buyer has a $70 budget. What is the greatest number of items the buyer can purchase? Set 18 + 4n <= 70. Subtract 18 to get 4n <= 52, then divide by 4: n <= 13. The greatest whole-number quantity is 13.

A common distractor is 17, found by dividing 70 by 4 and ignoring the fixed fee. Another is 13.5, which treats a fraction of an item as purchasable. The inequality makes the budget constraint explicit and the whole-number context resolves the final step. On the actual Quant section, you would do this without a calculator.

How to choose a target

Choose programs before deciding what score is enough. Compare the current class profiles and any school-specific minimums, then note whether your academic record makes a stronger Quant score especially useful. A target should be ambitious enough to support your application but tied to evidence from timed official-style practice. Avoid treating an average class score as a required threshold.

If your diagnostic is well below target, identify whether one section is limiting the total or whether all three need work. If one section is strong, preserve it with periodic practice while spending more time on the weakest section. If all scores fluctuate, build consistency before adding speed. After several comparable timed tests, use the trend and error log together rather than selecting the highest isolated result.

Turn a diagnostic into a study map

A diagnostic is most useful when you preserve the conditions and the reasoning behind its score. Take a full practice exam when rested, use the same section order you expect to use on test day, and take the optional break at the planned point. Then record total and section scores, the number of questions left unanswered, and question types that repeatedly consumed time. Avoid treating a single test as a precise forecast. Its purpose is to show where your next study block should go.

A practical log can use four columns: skill, mistake, correction and retest date. For example, under Data Insights you might write: 'Table Analysis; sorted by the wrong column; reread the condition before filtering; retest with a new table problem in two days.' That note is more useful than 'review DI.' Under Quant, 'forgot the fixed cost before dividing the budget' points to a setup error. Under Verbal, 'selected an answer that supports the premise but not the conclusion' names a reasoning gap.

Separate content errors from execution errors. A content error means you did not know how to manipulate an equation or assess an assumption. An execution error means you knew the method but misread 'least,' missed a negative sign, or copied a value incorrectly. Content needs instruction and drills. Execution needs a repeatable reading or checking habit. If you label every miss as careless, you may avoid repairing a genuine gap.

Schedule review so that new learning returns after a delay. A topic may feel clear immediately after watching an explanation, then disappear several days later. Revisit with a new problem without notes. If you can explain why each tempting distractor is wrong, you are more likely to transfer the rule to a changed question. If you can only recognize the same answer choice, keep practicing the underlying skill.

Build a realistic weekly schedule

A candidate with six available hours per week could use three two-hour blocks: one for Quant concepts and drills, one for Verbal or Data Insights, and one mixed timed set plus review. A candidate with ten hours could add a fourth focused block and a longer weekend session. The schedule should include actual solving time and review time. Counting passive video time as equivalent to active practice tends to overstate preparation.

Keep one session flexible for the weakest current skill. If a week's Quant review shows consistent accuracy but slow work with fractions, spend the flexible session on efficient fraction and ratio comparisons. If reading questions are accurate but slow, practice locating evidence and eliminating overbroad choices. A schedule that never changes despite diagnostic evidence is a calendar, not a learning plan.

At the end of each week, compare planned and completed work, then decide what to carry forward. Do not respond to a missed study block by doubling every session the next week. A sustainable plan with regular review is more likely to survive work and family obligations. If the timeline cannot accommodate both foundation work and several timed tests, moving the test date may be more useful than compressing all practice into the final days.

Registration decisions and retake room

The GMAT is offered throughout the year through online and test-center delivery, with appointments selected during registration. Year-round delivery is not the same as guaranteed access to any chosen time or location. When application deadlines are close, book early enough to leave room for results and a possible retake. A retake must be at least 16 calendar days after the earlier appointment, and results can take as long as 20 days.

Candidates may take the exam up to five times in a rolling 12-month period across both delivery methods, with no lifetime cap under the current policy. The policy also requires a five-year wait after a perfect score of 800 on the old scale or 805 on the current scale. Do not use the old lifetime cap quoted by outdated prep pages when planning attempts.

A retake should follow a change in preparation, not just a hope that a different test day will produce a better score. Before booking, identify what you will do differently: repair a specific algebra gap, improve reading pace, practice data sufficiency, or complete more full timed sections. Retaking immediately with the same habits gives little new evidence about whether the problem has changed.

Choose score sends deliberately

Candidates choose which programs receive official score reports. Five free sends are available within 48 hours after an official result becomes available. Make a shortlist before test day and confirm each program's score policy. If a program requires all attempts, that rule matters more than a general internet claim that candidates can hide a score. If a program accepts a superscore, understand which section results it will combine.

The GMAT Superscore uses the highest eligible section scores across valid attempts and appears with a valid score from one sitting. It has no separate percentile. Suppose a candidate improves Verbal on a second attempt but earns a lower Quant score than before. A superscore can preserve the earlier Quant result while using the stronger Verbal result. The program still decides whether that combined score is accepted, and the candidate should report according to its instructions.

Think about score sends and retakes as one application timeline. If the first result arrives above your target, you may choose to send it and focus on essays. If the result is below target and time remains, you can study the feedback and schedule another attempt, subject to spacing and appointment availability. There is no universal rule that every applicant benefits from retaking; weigh likely improvement against the time taken from other application work.

A short decision example

Maya has eight weeks before her first planned appointment and can study about eight hours each week. Her diagnostic shows strong Verbal performance, inconsistent Quant accuracy and the most severe timing problem in Data Insights. She can assign three weekly hours to Quant foundations, three to DI interpretation and timing, and two to Verbal maintenance with a timed set. After three weeks, a new diagnostic can show whether her DI bottleneck is calculator use, chart reading or question selection.

If the new test improves DI accuracy but leaves several answers blank, she should work on pacing and abandoning time-consuming items. If accuracy remains low even without time pressure, more timed work alone is unlikely to fix the issue. The next step is to classify the recurring content mistakes, learn the relevant methods and then add speed after accuracy stabilizes.

This example does not imply that eight weeks or eight hours per week is a universal recipe. It shows how a candidate can use available time, a diagnostic and a follow-up measurement to make a plan. Someone who has already studied these skills may need more timed practice. Someone returning to algebra after years away may need a longer foundation phase.