AP Calculus AB: The Complete Guide for 2026 & the May 2027 Exam

AP® Calculus AB is the most-taken AP mathematics exam and, for many students, the first real encounter with college-level mathematics. The College Board describes it as designed to be the equivalent of a first-semester college calculus course covering differential and integral calculus — roughly a college Calculus I. If your child took Algebra 1 in 8th grade, this is where the acceleration track usually lands.

A note on independence: AP® and Advanced Placement® are trademarks registered by the College Board. Texas CBE™ is an independent preparation platform, not affiliated with, authorized by, or endorsed by the College Board, which was not involved in producing this guide or our materials. We name AP® only to describe the exam we help students prepare for. We are not the exam, we do not reproduce College Board questions, scoring rubrics, or Course and Exam Description text, and we guarantee no particular score or outcome.

Start here: the exam format changes for May 2027

The May 2026 administration has already happened under the legacy format. The next exam most families are preparing for is May 2027 — and it is not the exam most prep sites describe.

The College Board has published a Clarifications and Corrections document for the AP Calculus AB and BC Course and Exam Description, flagged for implementation in Fall 2026. It revises the multiple-choice section in two places: Part A moves from 30 questions in 60 minutes to 29 questions in 62 minutes, and Part B from 15 questions in 45 minutes to 13 questions in 38 minutes. The AP Central course page states these updates take effect starting with the May 2027 exams.

Total exam length drops from 3 hours 15 minutes to 3 hours 10 minutes. The change lands entirely in Section I:

• Section I, Part A — no calculator: was 30 questions / 60 minutes → now 29 questions / 62 minutes (35% of the exam score).
• Section I, Part B — graphing calculator required: was 15 questions / 45 minutes → now 13 questions / 38 minutes (15% of the exam score).
• Section I total: was 45 questions / 105 minutes → now 42 questions / 100 minutes (50% of the exam score).

Section II is unchanged: Part A is 2 questions in 30 minutes, graphing calculator required (16.7% of the score); Part B is 4 questions in 60 minutes, no calculator (33.3%) — 6 questions, 90 minutes, 50% of the score.

The pacing implication is not a wash. Part A got slightly easier on the clock: one fewer question, two more minutes. Part B got tighter: two fewer questions but seven fewer minutes. Part B is the calculator section, home to the multi-step contextual problems, and now the most compressed part of Section I. Any practice set built on the old 45-minute Part B trains the wrong clock.

One practical note: the updated figures now appear in both the Course and Exam Description’s Exam Overview and the AP Central course page. Third-party prep material is uneven — much still describes the legacy 45-question section used through May 2026.

Exam mode. AP Calculus AB is a hybrid digital exam. Students answer the multiple-choice questions and view the free-response questions in the Bluebook application, but handwrite their free-response answers in paper booklets returned to the AP Program for scoring. Desmos graphing is built into Bluebook for calculator-permitted sections — practice in the tool rather than discovering it on exam day.

The calculator policy is inverted between sections — this is a real trap

Most students memorize “Part A, then Part B” and assume the calculator rule carries across the exam. It does not — it reverses between the two sections:

• Section I (multiple choice): Part A is no calculator; Part B requires a graphing calculator.
• Section II (free response): Part A requires a graphing calculator; Part B is no calculator.

The exam opens and closes with no calculator, calculator-active work in the middle — but the letters flip. A student who has internalized “Part B means calculator” will reach for a tool in free-response Part B they may not use. The drill: label every practice problem with the section and the calculator state. Free in October, expensive in May.

The eight units, and which ones actually decide the score

The course framework’s unit structure and exam weightings were not changed by this update. The Clarifications and Corrections document does revise the wording of two Essential Knowledge statements — one in Unit 5 on the Extreme Value Theorem, one in Unit 7 on differential equations — but those are clarifications of language, not new content.

A nuance almost every summary misses: the weighting table in the Course and Exam Description is explicitly titled as the exam weighting for the multiple-choice section. These percentages describe the multiple-choice section only — 50% of the score. The six free-response questions are not distributed by this table; they are built around modeling, graphical analysis, and justification tasks cutting across units.

The multiple-choice weightings for AP Calculus AB, as published in the Course and Exam Description:

• Unit 1 — Limits and Continuity: 10–15%
• Unit 2 — Differentiation: Definition and Fundamental Properties: 10–15%
• Unit 3 — Differentiation: Composite, Implicit, and Inverse Functions: 5–10%
• Unit 4 — Contextual Applications of Differentiation: 10–15%
• Unit 5 — Analytical Applications of Differentiation: 15–20%
• Unit 6 — Integration and Accumulation of Change: 15–20%
• Unit 7 — Differential Equations: 5–10%
• Unit 8 — Applications of Integration: 10–15%

A transparency note: the College Board’s student-facing course page publishes somewhat different bands — Unit 6 at 17–20%, Unit 5 at 15–18%, Units 1 and 2 at 10–12%, Unit 3 at 9–13%, Unit 7 at 6–12%. We publish the Course and Exam Description figures because that document is the authoritative framework and states its own scope. Either way, treat these bands as planning guides — no published weighting predicts an individual exam form.

What to take from this: Units 5 and 6 are the spine of the course — together roughly a third of the multiple-choice section under either set of bands, and where the free-response reasoning lives. Units 3 and 7 are lightest on multiple choice, which is not a reason to skip them: differential equations appear in free response, and Unit 3 supplies the chain rule everything after depends on.

Three big ideas recur throughout: Change, Limits, and Analysis of Functions. The framework puts sustained emphasis on communicating reasoning and justifications clearly — and the data below shows that is not decoration.

What the scores actually look like

The College Board has begun publishing 2026 results, but that release is still preliminary — rounded percentages only, without the student count, mean, or standard deviation. The preliminary 2026 AP Calculus AB distribution is 20% earning a 5, 28% a 4, 17% a 3, 24% a 2, and 11% a 1, with roughly 65% scoring 3 or higher. The most recent complete dataset is the May 2025 administration, and its shape is the story:

• Score of 5: 20.3% (58,174 students)
• Score of 4: 28.9% (82,970 students)
• Score of 3: 15.0% (42,896 students)
• Score of 2: 22.8% (65,405 students)
• Score of 1: 13.0% (37,277 students)
• 3 or higher: 64.2% — 184,040 of 286,722 students

Mean: 3.21. Standard deviation: 1.34. Roughly 36% scored below a 3.

That distribution is not a typo and not a bell curve. Almost twice as many students earned a 4 (28.9%) as earned a 3 (15.0%), and the 2 bucket is large (22.8%). AB tends to split into two populations — students who have internalized the reasoning and land at 4 or 5, and students who have not and land at 2 or 1 — with a thin middle. The 3 is the narrowest band, so there is no comfortable drift into one. Preliminary 2026 shows the same shape.

Where points are actually lost: the reasoning gap

This draws on a primary source almost no family reads: the Chief Reader Report, published each year by the person supervising the scoring of the free-response booklets. The 2025 AP Calculus report is credited to Sharon Taylor, Professor of Mathematics at Georgia Southern University, and records 2,201 readers across Calculus AB and BC.

The 2025 AB free-response mean scores, out of 9 points each: AB1: 3.26, AB2: 4.06, AB3: 4.82, AB4: 3.91, AB5: 3.60, AB6: 4.12.

The pattern is consistent: students are largely not losing points on algebra or arithmetic. They lose them on justification — on failing to say why. The examples below are our summaries of the report’s findings, not reproductions of College Board questions or scoring guidelines.

1. The global argument is the single biggest killer

On question AB1, part D asked students to determine where a modeled quantity attained its maximum. The point for supplying a global argument justifying the conclusion had a mean of 0.07 out of 1 — roughly 7% of responses earned it. On AB4, the corresponding justification point averaged 0.15 out of 1. The report attributes the shortfall to errors in attempted candidates tests, or to responses stopping short of a genuinely global argument by other approaches.

The underlying mistake is subtle and completely fixable. A response arguing only that the derivative changes sign from positive to negative at the critical point has made a local argument — it establishes a local maximum, not a maximum across the whole interval. To earn the point, a response must either carry out a full candidates test, comparing the value at the critical point against both endpoints, or state that the critical point is the only one on the interval, which promotes the local argument to a global one. Students find the right answer, then fail to prove it is the right answer. That is a writing habit, trainable in an afternoon.

2. Units on the wrong function

On AB3, students approximated the derivative of a reading-rate function by computing an average rate of change from a table. Because the function was measured in words per minute, its rate of change carries units of words per minute per minute. The report notes most responses supplying units gave correct ones — the more common problem was giving no units at all. Among those that did err, the main difficulty was supplying units for the original function rather than its derivative. The rule worth drilling: differentiating adds a second denominator, integrating a rate cancels it.

3. Asserting continuity instead of justifying it

Also on AB3, applying the Intermediate Value Theorem required first establishing continuity. The report observes that some responses simply asserted the function was continuous without justification. Of the nine points on that question, the continuity point and the point for correctly applying the theorem were earned by the smallest proportion of responses — means of 0.27 and 0.28 out of 1. The accepted move is one sentence: the function is differentiable, therefore continuous. Then check the hypotheses explicitly before invoking the conclusion. The report’s note to teachers is pointed — students apply the theorem’s conclusion without first checking or stating the conditions it requires.

4. Average value versus average rate of change

On AB1, the report identifies two leading misconceptions in part A: using the derivative of the function instead of the function itself as the integrand, and finding an average rate of change when the average value was asked for. Integrating a function’s derivative over an interval and dividing by the interval length yields its average rate of change — not the average value of the original function.

5. A bare calculator answer forfeits the setup point

On calculator-active free response, the setup earns a point independently of the final answer; writing only the final number throws that point away. A related trap appears on AB2, where the report notes an answer of 5.1 or 5.13 did not earn its point because of a decimal presentation error — the convention is three decimal places, rounded or truncated. Correct work with the wrong presentation still scores zero on that point.

Notice what is not on this list: computing derivatives, computing integrals, algebra. The reasoning gap — not the computation gap — separates the 4s from the 2s.

AB or BC? The honest comparison

This is the decision most acceleration-track families are really asking about, and it is usually framed wrong.

BC is a superset of AB. The College Board describes BC as applying the content and skills of AB to parametrically defined curves, polar curves, and vector-valued functions, while developing additional integration techniques and introducing sequences and series. AB is designed as the equivalent of a first-semester college calculus course; BC as the equivalent of both first and second semester.

The prerequisites are identical. The Course and Exam Description lists one set for both courses: the equivalent of four years of secondary mathematics for college-bound students, including algebra, geometry, trigonometry, analytic geometry, and elementary functions. BC does not demand a different kind of student — it is the same skills applied to more content, faster.

BC carries an AB subscore — a separate score reflecting performance on the AB-level portion. That is a genuine safety net: a student who struggles with series can still show a qualifying AB subscore.

The scores appear to favor BC — with a large caveat. In 2025, BC’s 3-or-higher rate was 78.6%, with 43.9% earning a 5, against AB’s 64.2%. The AB subscore earned by BC students averaged 4.11, with 88.0% at 3 or higher.

Do not misread those numbers. This is not evidence that BC is easier. BC’s cohort is self-selected — further along, often better supported, and smaller than AB’s (160,954 versus 286,722 in 2025). The gap reflects who takes it, not what it demands. A student moved from AB to BC does not inherit BC’s pass rate.

How to actually decide. The real questions are pace and foundation, not raw ability: how solid is the precalculus fluency — trig values without a lookup, algebra that does not consume working memory? What does the school actually offer, since a strong AB teacher beats a thin BC section? Is there room in the schedule? A reasonable default for the acceleration student with a strong precalculus foundation and schedule room: BC, with the AB subscore as insurance. For everyone else, AB done well — a 5 on AB beats a 3 on BC in essentially every respect.

Exam fees, and what Texas families actually pay

AP Exams are charged per subject, which surprises families planning a three-AP year. For the 2026 administration:

• Base fee at schools in the US, US territories, Canada, and DoDEA schools: $99 per exam.
• Exams at schools outside the US, US territories, Canada, and DoDEA schools: $129 per exam.
• College Board fee reduction for students with significant financial need: −$37 per exam.
• Texas Education Agency contribution for eligible fee-reduced public-school students: −$27 per exam.
• Resulting cost for an eligible, fee-reduced Texas public-school student: $26 per exam.
• Late-order fee, orders placed November 15 – March 13: +$40 per exam, on top of the exam fee. Unused or canceled exams also carry a $40 fee rather than the full amount.

That $99 to $26 path is underexposed among Texas families and worth a conversation with your AP coordinator. Two caveats matter:

• The Texas contribution is for public-school students only. Private-school and homeschool students may qualify for the College Board’s $37 reduction but not the state’s $27. Confirm with the AP coordinator at the school where your student will test.
• CEP and Title I enrollment is not automatic eligibility. The College Board is explicit that students at CEP and Title I schools and districts are not automatically eligible; the student must meet the fee-reduction criteria. Qualifying paths generally include the National School Lunch Program; SNAP, TANF, FDPIR, Medicaid, or Head Start; unhoused, migrant, or foster status; or family income within USDA guidelines.

Critically, your AP coordinator must indicate the student’s fee-reduction status in AP Registration and Ordering by a spring deadline — for 2026 it was April 30. Nobody does this for you. Ask in the fall, not in April.

When is the 2027 exam?

Honestly: it has not been published yet. The College Board typically releases the subject-by-subject schedule around the preceding fall, and the 2027 schedule is not out. For reference, the 2026 AP Calculus AB exam was administered Monday, May 11, 2026, at 8 a.m. local time, inside a regular window of May 4–8 and May 11–15, with late testing May 18–22. We will not invent a 2027 date — check the official AP Central exam dates page once it is published. A morning start and a two-week window in early-to-mid May is a reasonable planning assumption, but confirm before booking around it.

A realistic preparation timeline

The failure mode is not insufficient effort in April. It is April being the first time reasoning gets practiced under a clock. Over the summer, close precalculus gaps rather than starting calculus early — unit circle fluency, logarithm and exponential rules, automatic algebra. The Chief Reader Report explicitly encourages just-in-time review of precalculus topics inside the calculus course. From September, write justifications in full sentences the first week extrema appear: not “max at x = 3” but the complete argument, with sign analysis and endpoint comparison. The habit has to predate the pressure. In March, run full timed sections on the correct clock — for a May 2027 or later tester, Section I Part B at 13 questions in 38 minutes, not 15 in 45 — and rehearse the calculator inversion. In April, work released free response against model answers and keep an error log by type, not topic: “asserted without justifying,” “wrong units,” “no setup.” It will be short and repetitive. That is the point. The first week of May, taper — theorem hypotheses, the calculator map, the three-decimal convention. Not new content. Sleep.

College credit: check, never assume

AP credit policies are set by each institution, and frequently vary by department within one. A 3 is commonly described as a qualifying score, but many selective institutions and many calculus-gated majors — engineering, physics, some business and economics programs — require a 4 or 5. Some colleges grant credit, some placement without credit, some both, some neither. The same score can mean three different things at three schools on one application list.

Anyone who tells you “X colleges accept a 3” is quoting an aggregator, not a college. The only correct instruction: look up each college on the College Board’s AP Credit Policy Search, then confirm with that college’s registrar or department, because institutions revise these policies themselves. We make no promise about any college’s credit or placement policy, and neither should anyone selling you preparation.

How Texas CBE™ fits in

Our AP Calculus AB preparation targets the gap the Chief Reader data identifies. Students largely do not lose points because they cannot differentiate. They lose them because they will not write the global argument, they put units on the wrong function, they assert continuity instead of justifying it, and they hand in a bare calculator answer. Those are habits, and habits respond to forced active repetition. The details:

• A free practice quiz, no account required. Start there before spending anything — the fastest way to see whether the reasoning gap is your student’s actual problem.
• Regular price is $39.99 per AP course for six months of access. Independently authored, aligned to the published course framework, and modeled after the exam’s format — not a copy of it.
• Free response is self-scored against a model answer. To be clear: we do not grade free response. The student works the question, compares against a worked model answer, and scores their own response. For the failure modes above, honest self-scoring against a model is genuinely effective, and we would rather say exactly what it is than dress it up.
• Parents can see progress. With an accelerated student, nobody notices the drift until the score arrives.

We are an independent preparation platform. We are not the AP® Exam, we are not affiliated with or endorsed by the College Board, we do not reproduce College Board materials, and no preparation program — ours included — can guarantee a score.

Two things are worth doing this week. Look at your student’s last graded free response and ask: when they found a maximum, did they prove it was global — by comparing endpoints in a candidates test, or by establishing it was the only critical point? If not, you have found the highest-leverage fix available, and it is a writing fix. Second, if your student attends a Texas public school and meets the fee-reduction criteria, ask the AP coordinator now.

Figures reflect College Board publications reviewed in July 2026. Exam dates, fees, formats, and score distributions are revised regularly — verify against the official sources below before making decisions.