AP Score Curves Compared: How Many Points You Need for a 5 on Every Exam
AP scoring researchA side-by-side look at the estimated composite cutoffs on 21 AP exams, from Calculus BC at 57% for a 5 to Computer Science Principles at 83%.
AP score curves are the most misunderstood part of the exam. Across the 21 exams we publish researched cutoffs for, a 5 begins at roughly 57% of the composite on AP Calculus BC and roughly 83% on AP Computer Science Principles — a gap of 26 percentage points on the same 1–5 scale. This is the full comparison, plus an honest explanation of what a curve on an AP exam actually is.
A 26-point spread hides inside the phrase “the AP curve”
Every AP exam turns your raw multiple-choice and free-response points into a single composite score, then cuts that composite into five bands. The cut points are not the same from exam to exam, and they are not even close. On AP Calculus BC a 5 starts at about 57% of the composite. On AP Computer Science Principles it starts at about 83%.
Put that in student terms. On a 100-point composite, a Calculus BC student can drop 43 points and still collect a 5. A Computer Science Principles student can drop 17. Same score report, same colleges reading it, wildly different margin for error — and nothing in either course description warns you the gap is there.
That is why comparing AP score curves is worth doing before you lock in next year’s schedule, and why “my exam has a brutal curve” is usually the wrong conclusion to draw from one rough practice test. Below is every cutoff we have researched, side by side, followed by what is actually generating the differences.
What an AP curve actually is — and what it is not
The word “curve” comes from the classroom, where it normally means grading on a distribution: the teacher ranks the class, the top slice gets an A, and your grade depends on how everyone around you did. Most students assume the AP exam works the same way. It does not. There is no quota of 5s. If every student in the country cleared the composite threshold, every student would receive a 5, and College Board would publish that distribution without blinking.
What actually happens is equating. Every year’s exam is built to the same specification, but no two forms come out exactly equal in difficulty, so College Board uses statistical methods to place each new form on the same scale as previous ones and then sets the composite boundaries to match. The aim is stability of meaning: a 4 earned this year should represent the same level of achievement as a 4 earned three years ago.
Two consequences follow, and both run against intuition. First, a harder form pushes the cutoff down, not up. If this year’s free-response questions were unusually punishing, the composite required for a 5 falls to compensate — the adjustment answers the test, not the students. Second, the score distribution is an output, not an input. The 46% of Calculus BC students who earn 5s is the result of who sat the exam and how they did, never a target set in advance.
It also means the scale is anchored to something outside your cohort. College Board frames each score as a statement about readiness for the equivalent college course: a 3 is “qualified,” a 4 “well qualified,” a 5 “extremely well qualified.” Those labels, not a percentile rank, are what the boundaries exist to preserve. So the honest one-line definition of an AP curve is: the set of composite cut points that keeps a 5 meaning the same thing year after year.
College Board has never published an official raw-score-to-1–5 conversion for a live exam, and it re-equates every form every year. The composite percentages below are our best reconstructions from the public evidence available, and they move by a few points from administration to administration. Treat them as targets with a buffer, never as guarantees. The score distributions in the last column are different: those are official published figures.
AP score curves compared: all 21 exams
The table sorts every exam we cover from the most forgiving 5 threshold to the strictest. The first three columns are the estimated share of the composite you need for a 5, a 4 and a 3. The final column is the official percentage of students who actually earned a 5 — included because it disagrees with the cutoff far more often than students expect.
Each exam name links to its own calculator, where you can enter multiple-choice and free-response points and see the composite and predicted score for that specific exam rather than a generic percentage.
| Exam | Composite for a 5 | For a 4 | For a 3 | % earning a 5 |
|---|---|---|---|---|
| AP Calculus BC | 57% | 48% | 38% | 46% |
| AP Physics C: E&M | 58% | 43% | 35% | 24% |
| AP Physics C: Mechanics | 60% | 46% | 36% | 20% |
| AP Calculus AB | 64% | 53% | 42% | 20% |
| AP Human Geography | 68% | 56% | 46% | 19% |
| AP U.S. Government | 71% | 57% | 45% | 24% |
| AP Statistics | 71% | 56% | 44% | 17% |
| AP Environmental Science | 71% | 57% | 44% | 12.6% |
| AP Biology | 72% | 58% | 44% | 18.9% |
| AP Chemistry | 73% | 58% | 42% | 17.9% |
| AP Physics 1 | 74% | 56% | 42% | 19% |
| AP Psychology | 75% | 61% | 51% | 14.4% |
| AP English Literature | 75% | 65% | 53% | 16% |
| AP European History | 75% | 60% | 47% | 16% |
| AP World History | 78% | 58% | 50% | 13.9% |
| AP U.S. History | 78% | 59% | 48% | 14.2% |
| AP Computer Science A | 78% | 59% | 46% | 25.6% |
| AP Microeconomics | 79% | 62% | 51% | 19% |
| AP Macroeconomics | 79% | 66% | 57% | 19% |
| AP English Language | 80% | 69% | 58% | 13% |
| AP Computer Science Principles | 83% | 72% | 58% | 10% |
The cutoffs are our estimates. The 5 rates beside them are College Board’s own published figures, but they are not all drawn from the same administration: we refresh each exam page as new distributions are released, so some rows carry the latest year and others the previous year’s finals, which is why a few percentages show a decimal. Expect any of them to move a point or two. The current set always lives on College Board’s score distribution pages.
Why the cutoffs differ so much from exam to exam
Three things move a cutoff, and a fourth decides whether an exam can be placed on this axis at all. Only one of the four is what students mean when they say “difficulty,” which is why a low threshold is such a poor proxy for an easy exam.
- Equating the form. Each year’s questions are new, so College Board statistically links the new form to previous ones and shifts the composite boundaries so the standard holds. This is the only mechanism that responds to exam difficulty, and it typically moves a cutoff by a point or two, not ten.
- The shape of the raw-score scale. An exam built from a small number of long, multi-step free-response questions produces lower raw scores across the whole population than one built from many short, independently scored parts. Section weighting compounds this: when free response carries a large share of the composite and the rubric is unforgiving, the whole distribution of composites shifts down and the boundaries follow it down.
- Who sits the exam. Cutoffs are anchored to a standard, but the visible outcome — the share of students at each score — is driven by the cohort. A heavily self-selecting group of students produces a distribution that looks generous even when the boundary has not moved at all.
- Whether the exam has a composite at all. A course assessed through a research paper, a portfolio or a performance task has no multiple-choice-plus-free-response composite to convert, so it cannot be placed on this axis. That is a real limit on any curve comparison, including this one.
Cohort strength does more work than difficulty
The clearest evidence sits inside our own table. AP Physics C: Mechanics has the third most forgiving 5 threshold of the 21 exams here, at about 60% of the composite — yet only 20% of its students earn a 5. AP Calculus BC sits three points below it at 57%, and 46% of its students earn a 5. Two nearly identical curves, outcomes more than double apart. The curve is not what separates them; the exams and the students are.
Run it the other way and the same lesson appears. AP Statistics has a mid-range cutoff — 71% for a 5, 44% for a 3 — and still posts the lowest 3-or-higher rate of the 21 exams on this page, at 60%. Nothing about the boundary is harsh. Statistics is simply taken by a much broader mix of students than Calculus BC, many of them meeting formal probability for the first time.
AP Computer Science A is the strangest case. Its cutoff is strict, 78% of the composite for a 5, yet it still posts one of the highest 5 rates we cover at 25.6% — while 22.0% of the same cohort scores a 1. That bimodal shape reflects a population split between students who arrived already writing code and students meeting Java for the first time that year.
This is the honest reason a “generous curve” headline is misleading. Self-selecting cohorts in Calculus BC and both Physics C exams are why their pass rates look high, and a low boundary usually means points on that exam are genuinely hard to earn.
Which AP exams have the best curve?
If the question is which AP exams have the best curve in the narrow sense of “lowest composite needed for a 5,” our data gives a consistent top five: AP Calculus BC (57%), AP Physics C: E&M (58%), AP Physics C: Mechanics (60%), AP Calculus AB (64%) and AP Human Geography (68%).
Four of those five are calculus-level math and physics, which is not a coincidence. These exams ask a handful of long, multi-step questions where one early slip can cost most of a question, so raw scores sit low across the entire population. If the boundary for a 5 sat at 75% on Physics C, the 5 would nearly disappear. The forgiving cutoff is the system correcting for a punishing point scale — it is not a discount.
Two caveats keep this list honest. The first is coverage: these are the 21 exams we have researched cutoffs for, not the roughly 40 exams College Board scores. Look across the full roster and Calculus BC does not hold every record it holds here. In College Board’s most recent published distributions, AP Chinese Language (48% earning 5s) and AP Japanese Language (47%) both edge past its 46%, and several exams sit above its 82% pass rate — including AP Research (90% scoring 3 or higher), AP Seminar (88%), AP Chinese (85%), AP 2-D Art and Design (85%) and AP Spanish Language (83%). Calculus BC is best described as the most forgiving curve among the exams with a conventional composite, and the highest 5 rate of any non-language exam.
The second caveat is that several of those higher numbers are not curve comparisons at all. The world language exams report a total-group distribution that includes students with substantial exposure to the language outside the classroom: College Board reports that about 70% of AP Chinese test-takers are heritage speakers, and that about 72% of AP Spanish Language students and about 25% of AP French students had exposure to the language beyond the classroom. College Board publishes a separate standard-group distribution for exactly this reason. Meanwhile AP Seminar, AP Research, AP 2-D Art and Design, AP 3-D Art and Design and AP Drawing are portfolio and performance-task assessments with no composite to convert, so no cutoff percentage exists to compare. That is why our curve ranking stops at 21 exams even though College Board publishes score distributions for every exam.
A low cutoff tells you points are hard to earn on that exam, not that the exam is easy to pass. Read the boundary and the score distribution together, or you will draw the wrong conclusion from either one alone.
Turning a curve into an actual study target
A curve is only useful if it changes what you do on a Tuesday night. Converting a cutoff into a raw-point goal takes two minutes, and it is where an AP raw score to 5 question finally gets a concrete answer: not a percentage, but a number of multiple-choice questions and free-response points.
- Find your exam in the table above and note the composite percentage for your target score.
- Open that exam’s calculator and enter a realistic practice-test performance to see where your composite currently lands.
- Add a buffer of three to five composite points to your target, because the boundary is re-set every year and our figure is an estimate.
- Split the target across sections by weight. Section weighting differs by exam, so a weak free-response section can be offset far more easily on some exams than on others — only your exam’s own weights tell you by how much.
- Re-score practice free-response answers against the published scoring guidelines rather than your own generosity, then feed the honest points back in. Rubric-accurate self-scoring is the single biggest source of error in student predictions.
- Treat every cutoff as an estimate with a few points of drift, not a published table.
- Never read a low 5 threshold as evidence that an exam is easy — usually it means the opposite.
- Compare cutoffs only between exams that actually have a composite; portfolio and performance-task courses do not.
- Check the score distribution and the cutoff together; they answer different questions.
- Remember that pass rates reflect who signed up as much as how the exam was scored.
- Recheck the year on any distribution you cite — figures for the same exam can shift from one administration to the next.
Compare the curves on your own schedule
The practical use of comparing AP score curves is not picking the exam with the friendliest boundary — it is knowing, before you sit down, how much room for error you actually have. A student targeting a 5 on AP English Language (80% of the composite) is playing a different game from one targeting a 5 on AP Calculus BC (57%), and their practice plans should not look the same.
Every exam in the table has its own calculator with the full composite math, the estimated cutoffs and the current score distribution: browse them all in our AP exam score calculators, or start with the most searched ones — APUSH, AP Biology, AP Psychology, AP World History and AP Chemistry. For the official numbers straight from the source, College Board keeps its full set of score information for students online.
More scoring research is on the EduCalcDb blog, where we publish the reasoning behind every cutoff we estimate rather than asking you to take the number on faith.
Frequently Asked Questions
What are AP score curves?
“AP score curves” is shorthand for the composite-score boundaries that convert your raw points into a 1–5. Each exam scales multiple-choice and free-response points into a composite, and College Board sets four cut points on that composite each year through equating. The curve is those cut points. It is not a bell curve applied to your cohort, and there is no fixed quota of students at each score.
Which AP exams have the best curve?
Among the 21 exams we publish cutoffs for, AP Calculus BC has the most forgiving 5 threshold at roughly 57% of the composite, followed by AP Physics C: E&M at 58%, AP Physics C: Mechanics at 60%, AP Calculus AB at 64% and AP Human Geography at 68%. All five are estimates, and the low boundaries exist because raw scores on those exams run low — a forgiving curve is not an easy exam.
Which AP exam has the hardest curve?
Of the exams we cover, AP Computer Science Principles has the strictest estimated boundary: about 83% of the composite for a 5, and about 58% even for a 3. AP English Language is next at roughly 80%, with AP Macroeconomics and AP Microeconomics at about 79%.
What raw score do you need for a 5 on an AP exam?
There is no single number, because every exam has a different raw-point total and a different boundary. An AP raw score to 5 conversion always runs through the composite: raw multiple-choice and free-response points are weighted into a composite, and the composite is compared to a cut point — about 57% of it on AP Calculus BC, about 83% on AP Computer Science Principles. Enter your points in the calculator for your exam to see the raw totals behind those percentages.
Is the AP curve a bell curve? Do only a fixed percentage of students get 5s?
No. AP exams are not graded on a distribution and there is no cap on the number of 5s. Boundaries are set through equating so that a given score reflects the same level of achievement each year, which means the share of students earning each score is a result rather than a plan. In principle every student could earn a 5; in practice the distributions vary widely by exam because the cohorts do.
Why do AP cutoffs change every year?
Because the questions change every year. College Board writes a new form each administration and statistically links it to previous forms, then adjusts the composite boundaries so the standard stays constant. A slightly harder form gets a slightly lower cutoff and a slightly easier one gets a slightly higher cutoff. Movement is normally small — a point or two of composite — but it is enough that any published cutoff should be treated as an estimate.
Does a generous AP curve mean the exam is easy?
No, and this is the most common misreading in the data. AP Physics C: Mechanics needs only about 60% of the composite for a 5, yet just 20% of its students earn one. AP Calculus BC needs about 57% and 46% of its students earn a 5. A low boundary usually signals that points are hard to earn on that exam, and a high pass rate usually signals a strong, self-selecting group of test-takers.
Are these AP score cutoffs official?
No. College Board has never published an official raw-score-to-1–5 conversion for a live exam, and it re-equates every form annually, so every cutoff on this page and in our AP score calculators is an unofficial estimate. The score distributions are a different matter — those are official figures published by College Board for each exam.
Do different test dates or exam forms have different curves?
Slightly, and that is exactly what equating exists to handle. Students taking a late-testing or international administration sit different questions, and each form is placed on the same scale so that a 4 means the same thing whichever version you took. You cannot choose a test date to get a friendlier curve.