ReferenceNovember 10, 2026·7 min read

How Many Questions Can You Miss on the GMAT® and Still Score 655?

The GMAT® doesn't score by question count, but you can still plan with numbers. Here's why the count varies, the accuracy benchmarks that predict a 655, and the misses that deserve attention first.

Isaac Puglia
Isaac Puglia
Founder, The GMAT® Strategy · 99th percentile scorer

How Many Questions Can You Miss on the GMAT® and Still Score 655?

Turning a score goal into a question count is a smart instinct. A 655 is a number on a scale, and a number on a scale is hard to practice toward. Roughly 17 right answers out of 21 in the Quantitative Reasoning section is a target you can check yourself against this afternoon.

On a fixed-form test, that translation works: count the questions, weight the points, and the arithmetic tells you where you stand. The GMAT® needs one adjustment before the same arithmetic becomes useful, because the exam doesn't score by count. Your score reflects the difficulty level of the questions you're answering, so no exact count can guarantee a 655. Prep forums are full of tables that promise one anyway. Those tables are guesses.

The honest version of the answer has three parts: what the adaptive scoring system rewards, the accuracy benchmarks we use with students aiming for 655, and the measurement that estimates your level more directly than any count can.

Why the Count Isn't the Score

The GMAT® is question-adaptive. Answer a question correctly and the next one usually gets harder. Miss one and the next one usually gets easier. Each section is tracking the difficulty level of the questions you're seeing, and your score is roughly based on where you end up on that spectrum by the end of the section, not on a tally of correct answers.

That's what makes the same count mean different things. Imagine two test-takers who both answer 42 of the 64 questions correctly. One spent the section missing easy and medium questions, so the difficulty level of what they were seeing stayed low. The other missed only the hardest questions the exam could present, so they spent the whole section operating near the top of the difficulty range. Same count, different difficulty levels, different scores.

Section scores run from 60 to 90, and the total score combines them on a scale from 205 to 805, with each section weighted equally. A 655 total sits at the 90th percentile on the Focus Edition scale. The useful planning numbers are difficulty and accuracy, not a right-answer quota.

For the full mechanics, see our guide on how the GMAT® scoring algorithm works.

What the Count Can Tell You

One of our coaches reached the 99th percentile while answering about 65% of the questions correctly. That's roughly a third of the exam wrong, from a single test, so treat it as an illustration rather than a formula. It shows why a raw miss count doesn't predict a score on its own: the exam presents harder and harder questions until you start missing, which means missing is part of the design at any high difficulty level.

What mattered in that result was where the misses sat. The easy and medium questions were answered reliably, and the misses were concentrated among the harder questions.

That's the pattern behind a 655-level performance: reliable answers on easy and medium questions, with hard-question misses treated as expected rather than alarming.

In timed practice on official questions, the accuracy range we use as a TGS benchmark is roughly 70 to 80% before a 655 looks realistic. Translated into counts:

64×0.70=44.864×0.80=51.264 \times 0.70 = 44.8 \qquad 64 \times 0.80 = 51.2

About 45 to 51 of the 64 questions, as a practice benchmark. Split by section at the 80% mark:

21×0.80=16.823×0.80=18.420×0.80=1621 \times 0.80 = 16.8 \qquad 23 \times 0.80 = 18.4 \qquad 20 \times 0.80 = 16

About 17 of 21 in Quantitative Reasoning, 18 of 23 in Verbal Reasoning, and 16 of 20 in Data Insights. Those are planning numbers, not scoring rules. The exam doesn't tally correct answers.

What Matters More Than the Count

Which questions you miss

Three misses in a 21-question quant set may not tell you much on its own. Three easy misses usually tell you a lot. The Official Guide categorizes practice questions as easy, medium, or hard, so you can classify your own misses by difficulty instead of just tallying them. When a practice set shows repeated easy or medium misses, those deserve attention before isolated misses on the hardest questions, because repeated easy misses are usually what's holding the difficulty level down.

Whether the clock was running

An accuracy percentage from untimed practice usually overstates what you'd produce under section timing. The 70 to 80% benchmark applies to official questions attempted at pace, because pacing pressure is part of what causes mistakes. Build accuracy first, then check whether it holds when the clock is running. Our guide on how to get faster at GMAT® quant covers the speed-building side once accuracy holds.

Section balance

The three sections carry equal weight in your total, so a materially weaker section can limit the total even when the other two are strong. Comparing your section performance on an official mock, rather than adding up correct answers across the whole exam, is the more defensible way to decide where your study time goes. A quant section running five points below the other two usually deserves the attention before an overall accuracy average does.

How Many Can You Miss for a 700?

Searches for "how many questions wrong for a 700" are asking about the Classic 200-800 scale. GMAC's concordance table puts a Classic 700 at 655 on the Focus Edition, both at the 90th percentile. Same score, different scale, same answer.

The conversion table in our GMAT® score chart and percentiles guide covers the full Classic-to-Focus mapping if you're comparing against a school that still reports Classic numbers.

The Measurement That Beats Counting

The official practice exams from mba.com use the same scoring algorithm as the exam itself, which makes an official mock the most direct practical check on whether you're near 655 level. One mock can't guarantee a test-day result, but it's a better estimate than any count-based table. Third-party practice tests don't match the algorithm, so their scores can be less predictive of an official result.

Put together, the process looks like this: take an official mock for the measurement, use timed accuracy on official questions as the supporting benchmark, classify the misses by difficulty and section, and let the repeated easy and medium misses set your study priorities. Our guides on GMAT® practice test strategy and how to review GMAT® practice tests cover the cadence and the review system.

If the clock runs out with a question left blank, that can hurt your score more than a wrong answer would. In the final minute of any section, confirm an answer, even a guess.

Frequently Asked Questions

How many questions do you need to get right to score 655 on the GMAT®?

There's no fixed number. The GMAT® is question-adaptive, so your score reflects the difficulty level you reach, not a count of correct answers. As a practice benchmark, roughly 70 to 80% accuracy on official questions in timed conditions is the range we associate with 655-level performance, which comes out to about 45 to 51 of 64 questions.

How many questions can you get wrong and still score 655?

Plenty, if the misses are hard questions. One of our coaches reached the 99th percentile while answering about 65% of the questions correctly, which means missing roughly a third of the exam. At a 655 target, repeated easy and medium misses are usually the pattern that needs attention first, because those pull the difficulty level of your section down.

How many questions are on the GMAT®?

The GMAT® Focus Edition has 64 questions across three sections: 21 Quantitative Reasoning questions, 23 Verbal Reasoning questions, and 20 Data Insights questions. Each section runs 45 minutes, for 2 hours and 15 minutes of section time.

Does spending more time on a question improve your GMAT® score?

No. The scoring process includes whether you answered correctly and the difficulty of the question; time spent on an individual question isn't a scoring input. The indirect risk still exists, though: extra minutes on a hard question can leave later questions unanswered, and a blank final question can hurt more than a wrong answer.

What accuracy should you target in practice for a 655?

Roughly 70 to 80% on official questions in timed sets, with misses concentrated at the hard end. If you're in that range and repeated easy or medium misses aren't showing up, that's encouraging evidence, and the official mock score remains the better check.

How many questions can you miss and still score 700?

A Classic 700 corresponds to 655 on the Focus Edition scale, both at the 90th percentile, so the answer is the same as the 655 answer above. There's no fixed miss count; the difficulty of the questions you miss matters more than the number.

Want to learn even more?

Related reading:

Episode 56 of our podcast series, "What Does The GMAT® Scoring Algorithm Reward?", covers what the algorithm is designed to measure and how to study toward it. You can search for it on Spotify, Apple Podcasts, or YouTube, where you'll also find our full catalog of strategy episodes.

Want to learn even more?

Watch our free webinar on how to reach your dream GMAT® score in half the normal time — covers scoring, pacing, and the study approach that gets results fastest.

Or grab the free e-book — 3 keys to reaching your dream GMAT® score faster.