AP Calculus BC: The Complete 2026 Guide — Format Changes, Series, the AB Subscore & What Actually Costs Points

Read this first: the exam you are preparing for is not the exam described on most websites. College Board has published a clarifications and corrections document for the AP® Calculus AB and BC Course and Exam Description, marked for implementation in Fall 2026 — the 2026–27 course year. The multiple-choice section changed, and a great many third-party pages still describe the old one. Every number below was taken from College Board’s own published documents in July 2026.

Texas CBE™ is an independent preparation platform. We are not endorsed by, affiliated with, or connected to the College Board in any way. AP® and Advanced Placement® are registered trademarks of the College Board, used here only to identify the exam we help students prepare for. We do not reproduce College Board exam questions, scoring rubrics, or course description text.

This guide is for the family that has just registered for AP® Calculus BC — or is still deciding between BC and AB — and wants to know what the exam is, which units decide the score, where points are genuinely lost, and what to do about it starting Monday.

What AP Calculus BC actually is

The simplest honest description: BC is everything in AP Calculus AB at a faster pace, plus a body of material AB never touches. College Board describes AB as designed to be the equivalent of a first-semester college calculus course, and BC as the equivalent of both first- and second-semester college calculus.

The framework organizes the course into ten units. AB covers eight; BC covers all ten. But the unit count understates the gap, because BC also adds individual topics that sit inside the eight shared units and are labeled “bc only” in the framework. Those buried topics are the ones families overlook, because a unit list makes BC look like “AB plus two units at the end.” It is not. There are six BC-only topics hidden inside the shared units, by framework number and name:

• 6.11 — Integrating Using Integration by Parts
• 6.12 — Integrating Using Linear Partial Fractions
• 6.13 — Evaluating Improper Integrals
• 7.5 — Approximating Solutions Using Euler’s Method
• 7.9 — Logistic Models with Differential Equations
• 8.13 — The Arc Length of a Smooth, Planar Curve and Distance Traveled

Then come the two units AB students never see at all, which are BC-only from top to bottom:

• Unit 9: Parametric Equations, Polar Coordinates, and Vector-Valued Functions — nine topics, including arc lengths of parametric curves (9.3), vector-valued functions, and polar area
• Unit 10: Infinite Sequences and Series — convergence tests, Taylor and Maclaurin series, radius and interval of convergence, and error bounds

One trap in that list: Euler’s method is topic 7.5, not 7.7. Topics 7.6 and 7.7 are separation-of-variables topics that AB students also study. Third-party summaries get this wrong often enough to be worth checking any study list against the framework itself.

The exam format, as of July 2026

The exam is 3 hours and 10 minutes, with 42 multiple-choice questions and 6 free-response questions. The two sections are weighted equally at 50% each. Here is the structure as published:

What changed. Per College Board’s clarifications document, multiple-choice Part A went from 30 questions in 60 minutes to 29 questions in 62 minutes, and Part B went from 15 questions in 45 minutes to 13 questions in 38 minutes. These changes are marked for implementation in Fall 2026. If a book or website tells your student the first section has 45 multiple-choice questions and runs 1 hour 45 minutes, it is describing the earlier format.

The practical consequence is a pacing change that does not cut the same way in both parts. Part A now gives roughly 2 minutes 8 seconds per question — slightly more breathing room than the old 2 minutes flat. Part B gives roughly 2 minutes 55 seconds each, tighter than the old 3 minutes. Students drilling with old timing should re-time their calculator-active practice.

The administration is a hybrid digital exam: students complete the multiple-choice questions and view the free-response questions in College Board’s Bluebook testing app, but they handwrite their free-response answers in paper exam booklets that are returned for scoring. A student who has only ever practiced free-response on a screen has not practiced the thing they will actually do.

One more structural fact: AB and BC share three of the six free-response questions. The 2025 Chief Reader Report labels them AB1/BC1, AB3/BC3, and AB4/BC4 — the same prompts on both exams. Half of BC’s free-response section is, literally, the AB free-response section.

Exam dates: what is known and what is not

If you are reading this to plan a course starting in the fall, you are planning for a May 2027 administration — and as of July 2026, College Board has not published the 2027 exam calendar. The official calendar still covers the 2025–26 year only. We will not print a 2027 date we cannot source. AP Exams are historically administered across two weeks in May, with a late-testing window the following week. Check the official College Board calendar once the 2027 dates are posted, rather than trusting a date from a forum or a prep site.

The ten units and what they are worth

The framework publishes exam weighting for the multiple-choice section — an important qualifier that gets dropped constantly. These percentages describe Section I, not the whole exam. There is no published per-unit weighting for the free-response section.

Read the bottom two rows together. Units 9 and 10 — neither of which exists in AB — are 25–35% of the multiple-choice section between them. Add the six BC-only topics scattered through Units 6, 7, and 8, and it is clear why BC cannot be treated as a slightly brisker AB.

The AB subscore: the fact most parents have never been told

Every BC exam taker also receives a separate AP Calculus AB subscore on the same 1–5 scale. College Board describes it as reflecting the part of the exam devoted to Calculus AB topics, which it puts at “approximately 60% of the exam.” That approximate figure is the only public one; the exact weighting formula is not published, and you should be suspicious of any site claiming to know it.

College Board also states that it recommends “to colleges that they treat the Calculus AB subscore on the Calculus BC Exam the same way they would treat an AP Calculus AB Exam score, since common topics are tested at the same conceptual level in both Calculus AB and Calculus BC.” Whether a specific college follows that recommendation is entirely that college’s decision. Credit and placement policies vary by institution and they change without notice. We make no promise about what any college will grant; the only reliable source is the college’s own published policy. Check it before you count on it.

The 2025 Chief Reader Report publishes the actual subscore distribution — a table almost nobody cites:

In 2025, 88.0% of BC takers earned an AB subscore of 3 or higher, against 78.6% who earned a 3 or higher on the BC exam itself. 160,954 BC students were scored.

What parents should take from it: a BC student whose Units 9 and 10 collapse does not usually walk away with nothing. The AB-topic portion is reported separately, and it is typically strong. That is real, measurable downside protection. It is not a guarantee — nobody can promise your student any particular score — but it changes the risk calculation of choosing BC.

Series is the unit that decides BC scores

If you remember one sentence from this guide: Unit 10 is where BC scores are made and lost. Four independent facts converge on it.

It absorbs the most instructional time. Unit 10 carries roughly 17–18 suggested class periods for BC — more than any other unit in the course. (Unit 6 is next at about 15–16.) It holds fifteen topics, 10.1 through 10.15, and it is not one idea: it is a stack of convergence tests plus Taylor polynomials plus error bounds plus interval-of-convergence machinery, each with its own conditions.

It is tied for the heaviest weight. 15–20% of the multiple-choice section, level with Unit 6.

It is BC-only. No AB scaffolding, no earlier course that introduced it gently, no partial credit from prior knowledge. It is also conceptually unlike everything before it: most of calculus asks students to compute; series asks them to decide and justify — which test applies, whether its conditions are met, what that proves.

It is taught last, under time pressure. Unit 10 lands in spring, exactly when the calendar is thinnest and review has begun. A class running two weeks behind ends up teaching the most decision-heavy unit in the course in a rush.

The scoring data confirms it. On the 2025 exam, the free-response question on series (BC6) had a mean of 4.32 out of 9 — the second-lowest of the six BC questions. Within it, the point awarded for the analysis at the endpoints and the correct interval of convergence had a mean of 0.14 out of 1. Roughly one student in seven earned it.

Where BC students actually lose points

The 2025 Chief Reader Report publishes the mean score on every free-response question. This is the most useful public document about BC that almost no family reads:

The two worst questions were the two BC-only units. Not a coincidence — and the single best argument for how to allocate study time. The Chief Reader names the specific errors, and they are worth reading as a checklist, because they are teachable and most of them are not calculus mistakes at all:

On the series question

• Using a comparison test on an alternating series. The report is blunt: a comparison test “can only be used when the series being compared have positive terms.” This was named as a common error behind students not earning the endpoint point.
• Misidentifying a series. The report notes that many students claimed a series of the form 3/(n+1) “is the harmonic series” rather than showing that it diverges by comparison — direct or limit — to the harmonic series.
• Algebra that breaks the centering. Students simplified an inequality of the form −1 < (x−4)/3 < 1 into intervals not centered at x = 4. A Taylor series about x = 4 has an interval of convergence centered at 4. If your interval is not centered where the series is, you made an arithmetic error — a free self-check costing five seconds. The report attributes these to algebraic or arithmetic mistakes, not to a misunderstanding of series.
• Taking the limit as x → ∞ instead of n → ∞. In the ratio test, n is the index. x is not going anywhere.
• Omitting the absolute value on the ratio.
• Using 1 as the first term of a geometric series when it was not.
• Never stating a conclusion. Some responses, in the report’s words, “never gave a definitive answer,” presented in a way that made the conclusion difficult to determine. An answer a reader cannot find does not exist.

On the polar question

• Failing to square r(θ) in the area integral — described as the most common error on that part, and it cascaded: responses that lost the first point typically lost the two that followed, because those depended on the integrand being right.
• Presenting an answer with no setup. The most common reason for missing the very first point — one most students could compute — was simply not providing the setup.
• Multiplying by the wrong derivative in the rate-of-change part — the report lists dx/dt, dr/dt, and dθ/dr among the wrong choices.

The pattern underneath all of it

Look at the point-level means on the polar question. The point for justifying the answer — not computing it — had a mean of 0.06 out of 1. Fewer than one student in fifteen earned it. The Chief Reader describes responses giving “incorrect, incomplete, or insufficient global arguments,” and correct answers presented “with insufficient supporting work.”

That is the recurring theme of the whole report, and the thing families under-invest in: BC students lose more points to justification than to calculus. They can differentiate. They can integrate. They cannot reliably explain why a local observation proves a global claim — and the exam pays for the explanation, not the observation. The exam’s own directions say it plainly: answers without supporting work will usually not receive credit, and justifications require mathematical reasons and verification of the conditions under which a theorem or test applies.

How the calculator policy really works

A graphing calculator is required on multiple-choice Part B and free-response Part A, and not permitted on multiple-choice Part A and free-response Part B. Adding the published weights, roughly two-thirds of the exam’s weight (35% + 33.3%) is calculator-free. Two things follow, and both are counterintuitive:

On the no-calculator parts, the arithmetic is designed to be doable. Numbers are chosen to work out. A student facing ugly decimals on Part B of the free-response section has probably set the problem up wrong — that is a signal to re-read, not to grind.

On the calculator parts, the calculator is not the point. The Chief Reader’s comments on the calculator-active questions are almost entirely about setup and communication: answers presented without a setup, the wrong function used inside a correct-looking integral, points lost to decimal presentation errors. The report explicitly credits teachers, saying student responses showed “evidence that teachers are doing a great job guiding students in the use of calculators,” while noting that many responses would have benefited from a better understanding of which function to use for a derivative, an integral, or an evaluation. The machine is not the bottleneck — knowing which function to hand it is. Two habits pay for themselves: write the setup before touching the calculator, and observe the exam’s stated precision rule, which asks that final answers be accurate to three places after the decimal point unless otherwise specified.

BC or AB? Deciding honestly

Here are the 2025 results side by side, as published in the Chief Reader Report, and then the caveat that matters more than the numbers:

2025 is the most recent year for which College Board has published a full distribution with student counts and means, so it is the year we use.

Do not read this table as “BC is easier.” It is not, and reading it that way is how families get hurt. BC is a self-selected population: students who reach BC generally moved through algebra and precalculus early, often had stronger preparation, and chose the harder course on purpose. The exam is harder; the population is stronger; the stronger population wins by more than the exam takes away. Swap a marginal student into that population and the average does not follow them.

The honest asymmetry is the one above: BC carries the AB subscore as downside protection, and AB has no equivalent upside. A BC student who struggles with Units 9 and 10 still typically lands an AB subscore of 3 or higher — 88.0% did in 2025 — which College Board recommends colleges treat as an AB score, though each college decides for itself. An AB student who finds the course easy gets no BC credit for it.

So the practical question is not “which exam scores better?” It is: does this student have the time and the precalculus foundation to absorb Units 9 and 10 in the spring? Trigonometry that is not automatic makes polar painful. Algebra that is not clean destroys interval-of-convergence work. If the foundation is shaky, the fix is the foundation, not the course label.

What the exam costs

The figures below are College Board’s published fees for the 2025–26 school year. Amounts for 2026–27 were not published as of July 2026, so treat these as the current baseline rather than a promise about next year.

• Base fee: $99 per exam for exams taken in the U.S., U.S. territories, Canada, and DoDEA schools.
• College Board fee reduction: $37 per exam for eligible students with financial need. Schools are also expected to forgo their $9 rebate for fee-reduced students, so the arithmetic works out to $99 − $37 − $9 = $53 per exam.
• Texas: for qualifying students, the State of Texas contributes an additional $27 per AP Exam, which College Board states “brings the final cost to students down to $26.”
• Late-order fee: $40 per exam, in addition to the base exam fee.
• Unused or canceled exam fee: $40 per exam.

Those last two are the avoidable ones. Order on time and decide early — a student who drops the exam after the deadline still owes $40. Confirm current amounts, eligibility, and deadlines with your school’s AP coordinator; fees and dates are set annually and state funding is appropriated year by year.

A preparation timeline that matches how the course is actually taught

Fall (Units 1–6). AB territory plus the BC-only integration techniques — integration by parts, linear partial fractions, improper integrals. Do not let these feel optional because they are “just three topics.” They are the prerequisites for arc length, for improper-integral convergence arguments, and for much of series. Unit 6 is also 15–20% of the multiple-choice section on its own.

Winter (Units 7–9). Euler’s method, logistic models, arc length, then parametric and polar. Start free-response practice here: on paper, hand-written, timed — because that is exactly how the real thing works. Polar is a known killer — a 3.09 mean in 2025 — and far more forgiving in January than in April.

Late winter into spring (Unit 10). Series. Fifteen topics, the largest block of class time in the course. Here a student either builds a reliable decision procedure — look at the series, choose the test, check its conditions, state the conclusion — or memorizes a pile of tests and guesses under pressure. Build the procedure. Practice endpoints deliberately and separately, since that is where the 2025 mean collapsed to 0.14 out of 1.

April. Mixed sets under the new timing: 62 minutes for 29 no-calculator multiple-choice, 38 minutes for 13 calculator-active. Full free-response sets by hand. Re-read your own justifications and ask the reader’s question: would a stranger know what I concluded and why?

What to do Monday. Three things, in order. First, find last year’s BC free-response questions on AP Central and work BC2 and BC6 — the two hardest — untimed and honestly, then compare against the published scoring guidelines line by line. Second, write out from memory every convergence test with its conditions; if you cannot state the conditions, you cannot use the test — and the conditions are exactly what the Chief Reader says students get wrong. Third, take one problem you already solved correctly and write its justification as if to a skeptical stranger. That third exercise is the one nobody does, and it is worth the most points.

How we fit in

Texas CBE™ runs an AP® Exam Prep section built on the same active-recall engine as our other courses: you answer, you find out immediately, and the questions you get wrong come back until they stop being wrong. Our AP Calculus BC material is independently authored and organized against the ten published framework units, with the BC-only topics broken out rather than buried. We do not reproduce College Board questions, scoring rubrics, or course description text, and we are not the official exam or an official College Board product.

Two things we want to be straight about. Free-response practice on our platform is self-scored against a model answer — we do not grade free response, and no software should claim to replicate an AP Reader. It is still worth doing, precisely because the model answer shows what a justification looks like. And there is a free quiz that requires no account — try it before you spend anything. The regular price is $39.99 per AP course for six months of access.

We make no promise about your score and none about any college’s credit policy — nobody honestly can. But the evidence in this guide is real, it is public, and it points somewhere specific: Units 9 and 10, and the habit of justifying your answers. Whatever you use to prepare, spend your time there.