If you’re preparing for USMLE Step 1, you’ve probably seen students arguing about NBME 25, 26, 27, 28, 29, 30, 31, 32, and 33, and wondering which one you should actually sit for. Every forum thread seems to give a different answer.
Here’s the thing: “Which NBME is hardest?” is the wrong question. The questions that actually change your outcome are different ones. Which forms work well as a baseline? Which should you save for the last few weeks? How many do you really need? Should you take them online or offline? What does your score mean now that Step 1 is pass/fail? And what do you do if the number comes back low?
This guide answers all of those, in order.
NBME stands for the National Board of Medical Examiners. It’s the organization that writes the questions you’ll see on Step 1. For what those questions actually cover and how each system is weighted, see our breakdown of the Step 1 content outline.
When students say “I took an NBME,” they almost always mean the Comprehensive Basic Science Self-Assessment (CBSSA). That’s the official practice test built for Step 1. It has 200 multiple-choice questions split into four blocks of 50, and you can take it in standard timed mode or self-paced mode.
What makes a CBSSA different from any random question bank is what it gives you afterwards:
That estimated probability is the part students fixate on. Keep in mind that NBME presents it as an estimate based on your knowledge at that moment. It is not a guarantee of how you’ll perform on exam day.
An NBME self-assessment is not the same thing as the USMLE practice session at a Prometric test center.
The CBSSA is something you buy through the MyNBME portal and take from home or any suitable quiet location. The Prometric practice session is a separate product that lets you sit at an actual test center and work through USMLE sample materials in the real testing environment.
They serve different purposes. The CBSSA tells you where your knowledge stands. The Prometric session tells you what test day feels like. Most students take several CBSSAs and never book a Prometric session, and that’s fine. Just don’t assume one replaces the other.
There’s no magic number, and anyone who tells you “you must take seven” is making it up.
What matters is how much preparation time you have left and how much of the content you’ve actually covered. If you haven’t mapped that out yet, start with a study plan built to your timeline, then slot your NBMEs into it.
If you have plenty of time (8 weeks or more of dedicated study), take 5 to 7 forms, spaced out across your preparation. One every week and a half to two weeks works well for most people.
If you have moderate time (4 to 6 weeks), take 4 to 5 forms, leaning toward the newer ones. Spacing them roughly a week apart gives you enough room between tests to actually fix what you find.
If you’re short on time (under 3 weeks), take 3 to 4 forms. Don’t try to squeeze in every available test. Save your newest form for the final stage.
Every NBME costs money and takes a full day when you count the test plus the review. Taking eight forms and reviewing none of them is worse than taking three and reviewing all three properly.
The point of a self-assessment is not to collect scores. It’s to find out what you don’t know, fix it, and confirm that the fix worked. If you’re taking tests back to back without reviewing in between, you’re spending money to feel anxious.
A good rule: don’t schedule your next NBME until you’ve finished reviewing the last one and studied the weak areas it exposed.
If you want a straight answer, here it is.
Take an older form first. Save your newest forms for the end.
That’s the whole strategy. Here’s how it plays out across a typical preparation:
| Stage of preparation | Forms that can be used |
| Early / baseline | NBME 25-28 |
| Mid-preparation | NBME 29-31 |
| Late preparation | Your newest available forms |
| Final week | Free 120 |
You’ll see people claim older forms are “easier” or “outdated junk.” That’s not a useful way to think about it.
The real reason to save newer forms is timing. Newer forms are the most recent additions to the assessment pool, so they’re the closest reflection of how the test is being written right now. If you use them at week one, when you haven’t finished content review, you’ll burn your most informative test on a score that only tells you what you already knew: you’re not ready yet.
Older forms still contain solid Step 1 level questions. They’re perfectly good for finding out where you stand at the start and for tracking whether you’re improving. Use them for that.
A lot of students delay their baseline test because they’re scared of the number. Take it anyway, and take it early.
A baseline score is not a prediction. It’s a map. You’ll never get more useful information for one afternoon of work than a test that tells you exactly which systems are weak before you’ve committed your study plan. If you haven’t started content review at all, our Step 1 study guide for beginners covers what to do before that first form.
Students really want a difficulty ranking here. I’m not going to give you one, because it wouldn’t be honest. There is no official NBME ranking that declares one form harder, easier, or more predictive than the others. NBME describes these forms as tools for measuring readiness and tracking progress. It does not label any single form as the best one.
What’s actually useful is knowing when each form fits into your timeline.
| NBME form | Best use | When to take it |
| NBME 25 | Early assessment and review | Early |
| NBME 26 | Baseline or early progress check | Early |
| NBME 27 | Progress assessment | Early to mid |
| NBME 28 | Progress assessment | Early to mid |
| NBME 29 | Mid-preparation assessment | Mid |
| NBME 30 | Later progress check | Mid to late |
| NBME 31 | Late preparation | Late |
| NBME 32 | Recent assessment | Late |
| NBME 33 | Recent assessment | Final preparation |
NBME currently offers seven Comprehensive Basic Science Self-Assessment forms for purchase, with prices starting at $62 per form. Which seven are available changes over time, because NBME retires older forms and replaces them with new ones. NBME explains that forms are occasionally retired and replaced to provide a current, enhanced educational experience, and that some forms are retired at the same time new ones are released.
This matters for you practically. Higher form numbers that you see discussed in student groups or circulating as downloadable PDFs are not always forms you can buy and take with a real score report. Offline PDF copies give you questions. They do not give you a percent correct score, an estimated probability of passing, or content area feedback.
So before you build your schedule around specific numbers, log into MyNBME and look at what’s actually listed for purchase today. Plan around that, not around a forum post from two years ago.
Don’t build your study plan around a sentence like “NBME 31 is the most predictive form.” You’ll see this stated confidently on a lot of sites. There’s no official NBME statement backing it.
Short answer: there isn’t an officially established “most predictive” form, and the honest version of this answer is more useful than a fake ranking.
Here’s why rankings you find online disagree with each other.
Third-party datasets are not NBME data: Some websites and prep companies collect paired data from students who report both their practice scores and their real Step 1 outcome, then estimate which forms predicted best. That’s interesting, but it’s community-reported data with all the problems that come with it. Students who pass are more likely to report. Sample sizes vary enormously by form. And none of it is an official NBME prediction ranking.
Timing changes the answer: A form taken three days before your exam will always look “more predictive” than the same form taken six weeks out, because you were closer to your final knowledge level. That’s a fact about when you tested, not about the form.
Your own preparation shapes the result: If you took a form right after finishing cardiology, you’ll do better on a form weighted toward cardiology. Individual forms vary slightly in emphasis, and your personal strengths interact with that.
Your trend across several assessments, taken under timed conditions, in the last third of your preparation. That’s more informative than any single form, and it’s the thing you can actually control.
If your last three assessments all sit comfortably above the low-pass range and your estimated probability of passing is high on each, that’s a meaningful signal. If one form gave you a great number and the others didn’t, the great number is the outlier, not the truth.
Step 1 is pass/fail now. This changes how you should read your score, and a lot of advice online hasn’t caught up.
You no longer need a three-digit target. You don’t need to convert your NBME percentage into a predicted 240 or 250. That entire mental exercise is left over from the old scoring system, and chasing it will make you more anxious without making you more prepared.
NBME is explicit that the estimated probability of passing reflects your knowledge level at the time you tested, assuming you sit for Step 1 within about a week. It is not a promise.
Treat a high probability as good evidence, not as permission to stop studying. And treat a borderline probability as information worth acting on, whether that means more review time or a different study approach.
Both have a place. They just answer different questions.
This is the version that actually measures readiness. You get:
If you want to know whether you’re ready, this is what you buy.
Offline forms are cheaper or free and give you extra question exposure. They can be genuinely useful for learning, especially early on when you’re still building content knowledge and want more practice material.
What they don’t give you is a score you can trust, a probability estimate, or a content breakdown. You also lose the timed pressure, the interface, and the discipline of sitting through four blocks in one go.
Use offline forms to learn. Use online forms to assess. Don’t make a decision about your exam date based on an offline score.
This one is worth knowing if you’ve been studying for a while. In June 2026, NBME launched a new examinee interface for the Comprehensive Self-Assessments to align them more closely with the interface of the USMLE Step exams.
The new features include improved navigation for viewing questions and answer explanations, searchable lab values, a notes feature, text resizing up to 200%, image controls such as invert color, contrast and zoom, a dark theme option, and an improved highlighting tool.
If you last took an NBME before this change, the layout will look different. NBME also offers a free Sample Assessment on the MyNBME portal, under Purchase then Self-Assessments, with a tutorial and 12 sample questions laid out in the new interface. Spend fifteen minutes on that before your next timed form so you’re not learning the navigation during a test that’s supposed to be measuring your knowledge.
There’s no single correct date, because it depends on how long your preparation is. But the structure below works for most students.
3 to 6 weeks out: Take a newer form. This is your real readiness check, with enough time left to fix anything it exposes.
2 to 3 weeks out: Take another recent form. You’re confirming that the fixes worked and that the trend is holding.
Around 1 week out: Take your final recent assessment. This is the one closest to your actual knowledge level on test day.
Final few days: Free 120 and interface practice. At this point you’re not learning new content. You’re getting comfortable with question style and pacing.
Don’t take a full NBME the day before your exam. You’ll be tired, and if the score is lower than you hoped, you’ll walk into test day rattled. Nothing you learn that day will change your outcome.
Don’t leave a form untaken because you’re saving it. Students sometimes hoard their newest form out of anxiety and never use it. A form you didn’t take gave you zero information.
Student discussions frequently mention saving the newest forms and the Free 120 for the last stretch. That’s reasonable, and it’s what this guide recommends. Just be clear that these are student practices, not official NBME recommendations. NBME provides the tools and the score interpretation. It doesn’t publish a mandatory calendar.
This is the section most articles skip, and it’s the one that changes your score.
Taking the test is maybe 30% of the value. The review is the other 70%. Here’s how to do it without wasting a day.
1. Incorrect: You got it wrong. You don’t know the concept, or you know it incompletely. This is your primary study list.
2. Correct but guessed: You got it right and you know you weren’t sure. These are dangerous because your score says you know it, and you don’t. Treat these almost like incorrects.
3. Correct and confident: You knew it; you’re sure you knew it. Read the explanation quickly if you want, then move on. Don’t linger here. This is where students waste hours feeling productive.
Mark these as you take the test, not afterwards. Flag the ones you’re unsure about in the moment. Trying to remember your confidence level two days later doesn’t work. It’s the same method that makes question-bank review efficient, and our guide on how to review UWorld questions breaks down the six error types behind every miss.
After sorting, write down, on one page:
These categories need different fixes. A knowledge gap means study, and if it’s a concept that keeps slipping, make it an Anki card. A careless error means slow down and read the last line of the stem twice. A timing problem means practice pacing, not more content.
Don’t spend an entire day memorizing the answer key. You will never see these exact questions again.
The value of an NBME comes from understanding why you missed a question, not from remembering that the answer to question 47 was C. If you can explain the concept behind a missed question to someone else, you’ve reviewed it properly. If you can only remember the answer, you haven’t.
For most students, a few hours spread over one to two days after the test. If it’s taking you three full days, you’re reviewing too much of what you already knew.
If the same systems keep coming back weak across forms, that’s a content problem, not a review problem. Our Step 1 crash courses rebuild those systems concept-first.
Yes, with limits.
NBME periodically retires forms and replaces them with new versions. The stated reason is to keep the assessment experience current and aligned with the exam. Retired forms aren’t “wrong”; they’re just no longer the best mirror of what you’re about to sit for.
Use them as practice material. Use current forms as your measuring stick.
These get compared as if one has to win. They don’t do the same job.
| Resource | Main purpose |
| NBME CBSSA | Readiness check plus diagnostic feedback from the people who write Step 1 |
| UWorld QBank | Learning and question practice, with detailed explanations |
| UWorld Self-Assessment (UWSA) | An additional timed self-assessment |
| Free 120 | Familiarity with official USMLE-style questions and interface |
UWorld QBank is a teaching tool. It’s where you build knowledge. Doing thousands of UWorld questions is study, not assessment, and your UWorld percentage is not a readiness score.
UWSA gives you another timed data point. Useful as a supplement. Some students find their UWSA scores run differently from their NBME scores, which is normal since they’re different products from different organizations.
Free 120 isn’t really a readiness test. It’s official USMLE sample material, and its value is showing you what real question style and interface feel like. Take it close to your exam for familiarity, not to generate a number.
CBSSA is the readiness tool. It’s the only one on this list that gives you an estimated probability of passing Step 1.
The mistake to avoid is treating them as interchangeable scores. A strong UWorld percentage does not substitute for a timed CBSSA, and a Free 120 result doesn’t mean much as a score.
First: one low score is not a reason to panic, and it’s not automatically a reason to postpone your exam.
Before you make any decision, work through these questions honestly.
Was this your first assessment? Baseline scores are supposed to be low. That’s the entire point of a baseline.
Have you finished content review? If you still have three systems you haven’t touched, your score is telling you about unfinished work, not about your ability.
Did you run out of time? A timing problem looks like a knowledge problem on the score report. Check whether your errors cluster at the end of blocks.
Were mistakes concentrated in particular systems? Scattered errors across everything mean broad gaps. Errors concentrated in two systems mean a focused fix, which is much better news. Weight that fix by what the exam actually rewards, using our list of which topics carry the most question weight.
Are your scores improving? A low score that’s higher than your last low score is progress. Plot the trend.
What does the estimated probability of passing say? This is the most directly relevant number on the report. Look at it before you look at anything else.
Did you test under real conditions? Self-paced at midnight after a full study day is not a fair measure of you.
A single assessment is one data point. A trend across multiple assessments is what should inform your planning.
If you’ve taken three recent forms under timed conditions, all in the same disappointing range, with content review complete, that’s real information and worth acting on. If you’ve taken one form, early, while tired, with two systems unstudied, it isn’t.
And if you’re an IMG weighing a delay against application timelines, that’s a separate calculation involving your ERAS deadlines and your overall application. Make that decision with the full picture, not with one score report open at 2 am.
For IMGs, your Step 1 date touches Step 2 CK, USCE and ERAS. Book a free call and we’ll weigh your NBME trend against your full Match timeline.
Adapt these to what’s actually available for purchase on MyNBME when you’re planning.
Space these roughly every 10 to 14 days. Study in between. Don’t stack them.
Roughly one per week, with review days built in.
Don’t try to complete every form you own. With limited time, studying your weak areas will do more for you than generating three more scores you have no time to act on.
Earlier forms build and check your baseline. Newer forms confirm readiness. Every test gets reviewed. Your decision comes from the trend, not from any single number.
There isn’t one universally best form. NBME doesn’t designate a single form as the best or most predictive. The most useful approach is to use earlier forms to build your baseline, save newer forms for later readiness checks, and judge your readiness on the trend across several assessments.
Start with one of the older available forms, typically NBME 25 or 26. Save your newest forms for the final weeks, when they’ll actually reflect your test-day knowledge level.
Between 3 and 7, depending on how much time you have. With 8 or more weeks, 5 to 7 is reasonable. With under 3 weeks, 3 to 4. Taking more forms than you have time to review properly doesn’t help.
Yes, as a mid-to-late preparation assessment. There’s nothing wrong with it. It’s a solid progress check once you’ve finished most of your content review.
It can be, taken at the right time, but no official NBME data designates it as the most predictive form. Any form taken under timed conditions close to your exam date will be more informative than one taken early in your preparation.
Check MyNBME to see which forms are actually available for purchase before planning around specific numbers. Available forms change as NBME retires older ones and releases new ones. If a high-numbered form is available to buy, use it as one of your late assessments. If you’re only finding it as a downloadable PDF, treat it as practice questions rather than a readiness test, since offline copies don’t come with a score report or probability estimate.
There isn’t a fixed passing score on a CBSSA, because it isn’t a pass/fail test. Your score report gives you an estimated probability of passing Step 1 and shows where your score sits relative to the low-pass range. Use those rather than looking for a cutoff number.
Yes, for question exposure and concept learning. No, for judging final readiness or for score conversion. Retired forms weren’t written against the current content distribution, and old conversion tables were built for a scored Step 1 that no longer exists.
Take them online when you want to measure readiness, since that’s the only way to get a real score report, a probability estimate, and content area feedback under timed conditions. Offline copies are fine for extra practice while you’re still learning content.
They feel different rather than harder or easier. UWorld questions tend to be longer and more teaching-oriented, since the product is built for learning. NBME questions often feel shorter and more direct. Many students find the shift jarring the first time. That’s one more reason to take a full NBME before test day instead of relying only on UWorld.
Around a week before your exam for most students. Close enough to reflect your real knowledge level, far enough out that a disappointing number doesn’t wreck your confidence the night before.
Yes. Free 120 works best in the final few days, after your last full assessment. Its value is familiarity with official question style and interface rather than generating a score.
There isn’t one best NBME for Step 1.
Use earlier forms to establish and build your baseline. Save newer forms for later readiness checks. Review every single assessment properly, using the three-bucket method rather than memorizing answers. Then decide how ready you are based on your performance trend across several assessments, not on any single score.
That approach will serve you better than any ranking of NBME 30 against NBME 31. And once Step 1 is behind you, the exam that carries a three-digit score is Step 2 CK, our guide on scoring 250+ on Step 2 CK as an IMG picks up where this one ends.
Planning your Step 1 around your Match timeline? IMG Helping Hands breaks down the full IMG pathway, from USMLE scheduling to ERAS submission. Explore our IMG residency guides.
A score report tells you where you are. Deciding what to do next, with your ERAS timeline running in the background, is where most IMGs get stuck. Here is where we come in:
Articles published by IMG Helping Hands are prepared by our team using information from direct experience, publicly available resources, and educational references. AI tools may be used to assist with drafting, proofreading, and formatting; however, all content undergoes review and approval before publication.
The information provided is intended for educational purposes only. Requirements, policies, and processes may change over time. Readers should consult official sources for the most current information.
Get our most valuable tips right inside your inbox, once per month!
