AP Exam Playbook
How the May 2027 AP Calculus BC exam works, how free-response answers earn points, what your calculator has to do, what the scores mean (including the AB subscore), and a study plan that counts down to exam day.
Exam countdown
180days to go
Exam day: Monday, May 10, 2027, about 25 weeks away.
How the exam works
The AP Calculus BC exam is on Monday, May 10, 2027, in the morning session. That's the same time as AP Calculus AB, so you can't take both in the same year. At most schools in the U.S. it starts at 8 a.m. local time; some places use a different time, so check with your AP coordinator.
The exam takes 3 hours and 10 minutes and has 48 questions: 42 multiple-choice and 6 free-response. The two sections count equally. Multiple choice is half your score, and free response is the other half.
Swipe the table sideways to see every column.
| Part | Questions | Time | Calculator | Share of score |
|---|---|---|---|---|
| Section I, Part A: multiple choice | 29 | 62 minutes | Not allowed | 35% |
| Section I, Part B: multiple choice | 13 | 38 minutes | Graphing calculator required | 15% |
| Section II, Part A: free response | 2 | 30 minutes | Graphing calculator required | 16.7% |
| Section II, Part B: free response | 4 | 60 minutes | Not allowed | 33.3% |
| Whole exam | 48 | 3 hours 10 minutes | In two of the four parts | 100% |
That's a little over 2 minutes per question in multiple-choice Part A, a little under 3 minutes per question in Part B, and about 15 minutes per free-response question.
What's on it
BC covers everything in AB and more. About 60% of the exam is on the topics the two courses share, and the other 40% or so is BC only. In this course, Units 1 to 8 are the AB units with the BC-only topics added (integration by parts, partial fractions, improper integrals, Euler's method, logistic growth and arc length), and Units 9 and 10 are BC only: parametric, polar and vector functions, and infinite series. Those two units alone are 25% to 35% of the multiple-choice section, and series had a free-response question on every exam from 2023 to 2026.
Expect every kind of function you've studied, as formulas, graphs, tables and descriptions in words, plus the BC kinds: parametric and polar curves, vector-valued functions and power series. In multiple-choice Part B, not every question needs the calculator. Of the six free-response questions, at least two are set in a real-world situation, and about half of the work is carrying out procedures while the other half is explaining and reasoning.
The six free-response questions
Three of the six free-response questions are the same as on the AB exam, and three are BC only. On every exam from 2023 to 2026 they fell into the same slots:
| Question | What it's been, 2023 to 2026 |
|---|---|
| Q1 (calculator) | Shared with AB: rates and amounts in a real-world setting. |
| Q2 (calculator) | BC only: a particle moving in the plane (parametric or vector motion) in 2023 and 2024, a polar curve in 2025 and 2026. |
| Q3 (no calculator) | Shared with AB: a differential equation, or a model built from a table. |
| Q4 (no calculator) | Shared with AB: the graph of f′ and an accumulation function. |
| Q5 (no calculator) | BC only: a mix of BC techniques, such as integration by parts, an improper integral, arc length, Euler's method, or a Taylor polynomial from a differential equation. |
| Q6 (no calculator) | BC only, and always series: Taylor polynomials and series, an interval of convergence, an error bound. |
That's a pattern, not a promise: the College Board says three questions are shared with AB, but the slots could change. The BC-only questions are worth 27 of the 54 free-response points, and they often open with a part that's plain AB calculus, so start every one of them.
Screen for multiple choice, paper for free response
The exam is hybrid digital. Here's what that means in practice:
- You answer the multiple-choice questions in the Bluebook app. Each one has four answer choices, A to D.
- The free-response questions appear in Bluebook, but you write your answers by hand in a paper booklet, in pencil or in black or dark-blue pen. Only what's in that booklet is scored. Scratch paper isn't.
- Each part has its own time limit. Within a part, you can move between its questions until time runs out.
- The on-screen clock turns red when 5 minutes are left. Nobody announces it, so glance at the clock now and then.
- There's no formula sheet. The derivatives, integrals and theorems you need have to come from memory.
What to bring
- Up to two graphing calculators from the College Board's approved list, with fresh batteries or a full charge. (Bluebook's built-in Desmos calculator is there too; see Calculator skills below.)
- Pencils, or black or dark-blue pens, for the free-response booklet.
- Anything else your AP coordinator asks for. They confirm the start time, the room and how Bluebook is set up at your school.
Before May, try a Bluebook test preview so the app and its built-in calculator feel familiar, including graphing a parametric and a polar curve.
No formula sheet: what to know by heart
With no formula sheet, BC has more to remember than AB. Know these cold, along with the series tests in the free-response section below:
| What | Formula |
|---|---|
| Geometric series | a + ar + ar² + … = a/(1 − r), when |r| < 1 |
| Taylor polynomial about x = a | f(a) + f′(a)(x − a) + f″(a)(x − a)²/2! + … + f⁽ⁿ⁾(a)(x − a)ⁿ/n! |
| eˣ | 1 + x + x²/2! + x³/3! + …, for every x |
| sin x | x − x³/3! + x⁵/5! − …, for every x |
| cos x | 1 − x²/2! + x⁴/4! − …, for every x |
| 1/(1 − x) | 1 + x + x² + x³ + …, for |x| < 1 |
| Alternating series error bound | |error| ≤ the size of the first term you left out |
| Lagrange error bound | |f(x) − Pₙ(x)| ≤ (max of |f⁽ⁿ⁺¹⁾(z)| for z between a and x) · |x − a|ⁿ⁺¹/(n + 1)! |
| Integration by parts | ∫ u dv = uv − ∫ v du |
| Arc length of y = f(x) | ∫ₐᵇ √(1 + (f′(x))²) dx |
| Speed in the plane | √((x′(t))² + (y′(t))²) |
| Distance traveled in the plane (arc length of the path) | ∫ₐᵇ √((x′(t))² + (y′(t))²) dt |
| Parametric slopes | dy/dx = (dy/dt)/(dx/dt), and d²y/dx² = (the t-derivative of dy/dx)/(dx/dt) |
| Polar coordinates | x = r cos θ, y = r sin θ |
| Polar area | ½ ∫ r² dθ, between two angles |
| Euler's method | new y = old y + Δx · (dy/dx at the current point) |
| Logistic growth | dy/dt = ky(a − y): y heads to a, the carrying capacity, and grows fastest at y = a/2 |
How free-response questions are graded (and how to avoid losing points)
Each of the 6 free-response questions is worth 9 points, 54 points in all, and together they're half your score. A question usually has 3 or 4 lettered parts worth 1 to 3 points each.
The graders (the College Board calls them readers) score each point on its own. Since 2025, the scoring guidelines number the points in every question, P1 to P9, and give each one a short name for what it rewards: a setup, an answer, a reason. That's good news for you. A question you can't finish can still earn most of its points, as long as you write down every piece you can.
The kinds of points
| Kind of point | What earns it |
|---|---|
| Setup | The right integral, derivative, equation or formula, written in math notation, with the limits on a definite integral. |
| Form | The right structure: a Riemann sum built from the right widths and heights, or a volume integrand with each radius written out. |
| Answer | The correct value. On calculator parts this point often counts with or without work, but the setup point doesn't. |
| Answer with supporting work | The value plus the work that produces it, such as the difference and the quotient behind a slope estimated from a table. |
| Key step | Showing the central idea, such as setting f′(x) = 0 or looking at the sign of the second derivative. |
| Antiderivative | A correct antiderivative, including the right coefficient. |
| Units | The correct units, when the question asks for them. You can earn this point even if the number is wrong. |
| Interpretation | What a value means in the situation: the quantity, its units, and the time or interval. |
| Reason or justification | The right conclusion with a reason based on the information given, including the conditions of any theorem you use. |
Points are often linked. A reason point usually needs the correct answer it supports, and some points are only available if an earlier one was earned. In a separation-of-variables part, for example, the points usually build on each other: if you never separate the variables, none of them can be earned, and leaving out the constant of integration limits how many you can get. One bad first step can cost a whole part, so slow down at the start.
Show the work: no bald answers
A bald answer is a value with nothing behind it. On a no-calculator part it usually earns the answer point at most, and sometimes nothing at all: an unsupported Riemann-sum value, or "yes, by the IVT" with no values, can score zero. If a part says justify, explain or show the work, the work is what's being scored.
Loses the point
f′(3) ≈ 3
Earns the point
f′(3) ≈ (f(5) − f(2)) / (5 − 2) = (16 − 7) / 3 = 3
The number is right, but nothing shows where it came from.
You don't have to simplify. Once you've written a correct numerical expression you can usually stop, and a later arithmetic slip often won't cost the answer point. But a few parts do count errors made while simplifying, and an expression a reader can't read one way only (unmatched parentheses, for instance) earns nothing. Write the clean expression, then simplify only if it's easy.
Justify with calculus, and check the conditions
A justification is a reason built from the information you're given: the sign of a derivative, the values of a function, a theorem whose conditions you've checked. Name the function every time. "It", "the function", "the graph" or "the slope" with no name attached don't earn reason points. A sign chart on its own doesn't either: say in words or symbols what the sign is, on which interval, and what that tells you.
| To show that… | Say this |
|---|---|
| f is increasing (decreasing) on an interval | f′(x) > 0 (or < 0) on that interval. |
| f has a relative maximum (minimum) at x = c | f′ changes from positive to negative (negative to positive) at x = c. Or: f′(c) = 0 and f″(c) < 0 (> 0). |
| f has an absolute maximum or minimum on [a, b] | Evaluate f at every critical point and at both endpoints, then name the largest or smallest. A first or second derivative test on its own only shows a relative extremum. |
| The graph of f is concave up (down) | f″(x) > 0 (< 0) there. If you're given the graph of f′: f′ is increasing (decreasing) there. |
| f has a point of inflection at x = c | f″ changes sign at x = c. From a graph of f′: f′ changes from increasing to decreasing, or the reverse, at x = c. |
| f is increasing and concave down | Both conditions: f′ > 0 and f′ is decreasing on the interval. |
| A particle is speeding up (slowing down) at t = c | v(c) and a(c) have the same sign (opposite signs). State both signs. |
| A particle changes direction at t = c | v(t) changes sign at t = c. Saying v(c) = 0 isn't enough. |
| f(c) = k for some c (Intermediate Value Theorem) | f is continuous on [a, b], with a reason (for example, f is differentiable). Give f(a) and f(b), say k is between them, then conclude. |
| f′(c) equals the average rate of change (Mean Value Theorem) | f is continuous on [a, b] and differentiable on (a, b). Compute (f(b) − f(a)) / (b − a), then conclude f′(c) equals it for some c in (a, b). |
| A tangent-line estimate is too big (too small) | The graph is concave down (up) on the whole interval between the point of tangency and the point you're estimating. |
| L'Hospital's Rule applies | Show that the numerator and the denominator both go to 0 (or both to ±∞), then take the limit of the ratio of their derivatives. |
Loses the point
f has a relative maximum at x = 3 because f′(3) = 0.
Earns the point
f has a relative maximum at x = 3 because f′(x) changes from positive to negative at x = 3.
f′(3) = 0 on its own doesn't say what kind of point x = 3 is. And "f changes from increasing to decreasing" only restates what a maximum is: the reason has to come from f′.
Loses the point
f is concave down on (1, 4) because f′ < 0 there.
Earns the point
f is concave down on (1, 4) because f′ is decreasing there.
Concavity comes from how f′ is changing, not from its sign.
Loses the point
The particle is speeding up at t = 2 because a(2) > 0.
Earns the point
v(2) > 0 and a(2) > 0 have the same sign, so the particle is speeding up at t = 2.
Acceleration alone doesn't decide it: with negative velocity, a positive acceleration means slowing down.
Loses the point
Yes, by the IVT.
Earns the point
h is differentiable, so h is continuous on [1, 5]. h(1) = 2 < 6 < 9 = h(5), so by the Intermediate Value Theorem h(c) = 6 for some c in (1, 5).
Give the values, show the target is between them, and say why the function is continuous.
Loses the point
g′ changes from positive to negative at x = 2, so g has its absolute maximum at x = 2.
Earns the point
Compare the candidates: g(0) = 1, g(2) = 5, g(5) = 2 and g(6) = 3. The largest is g(2) = 5, so the absolute maximum of g on [0, 6] is at x = 2.
A sign change at x = 2 only shows a relative maximum. An absolute maximum needs a comparison over the whole interval: every critical point and both endpoints.
You don't have to name a theorem if its conditions and conclusion are clear. But if you name one, it has to be the right one. Mixing up the Intermediate Value Theorem (about values of f) and the Mean Value Theorem (about values of f′) is a classic way to lose a point.
Notation and linkage
- Use = only between things that are equal. A chain like f′(x) = 2x = 6 claims that 2x is always 6; write f′(3) = 2(3) = 6 instead. Errors like this come up every year.
- Write a definite integral in full: the integrand in parentheses, the limits, and dx (or dt). Never put ∫ in front of an antiderivative you've already found.
- Use ≈ for estimates, such as Riemann sums and tangent-line approximations.
- Keep dy/dx and dy/dt straight, and don't write a bare dy.
- Write lim in front of every expression in a chain of limits, including the one after you use L'Hospital's Rule.
- Use the names you're given. ∫₁⁴ (f(x) − g(x)) dx is enough, and it's safer than substituting long formulas, where parentheses get dropped.
- Label each part (a), (b) and so on, and label any function, graph or table you use.
- Copy given functions carefully. Re-read the formula before you differentiate it.
- Answer exactly what's asked. If the question asks for x-values, give x-values, not points. An extra wrong answer, like one inflection point too many, can cost both the answer and the reason.
- Extra work has to be right too. A wrong value you didn't need can still cost a point.
Loses the point
f′(x) = 2x = 6
Earns the point
f′(x) = 2x, so f′(3) = 2(3) = 6
Loses the point
∫ 3t² − 2t = t³ − t² = 4
Earns the point
∫₀² (3t² − 2t) dt = (2³ − 2²) − (0³ − 0²) = 4
The first line has no limits, no dt and no parentheses, and it says one thing equals both a function and a number.
Decimals: three places, and no early rounding
- Give decimal answers correct to three places after the decimal point, unless the question asks for something else (like the nearest whole number). Rounding and truncating both work: 1.29583… can be written 1.296 or 1.295. 1.3 and 1.30 aren't accurate enough.
- Within one question you lose at most one point for rounding. It's still a point, so don't give it away.
- Don't round in the middle. Store intermediate values in your calculator and use the stored value in the next step.
- Any intermediate value you write down, like the place where two curves cross, has to be correct to three decimal places itself.
Loses the point
a ≈ 0.79, so ∫₀ᵃ eˣ dx = 1.203
Earns the point
a = 0.792 (full value stored in the calculator), so ∫₀ᵃ eˣ dx = 1.208
Rounding the crossing point to two places moves the answer in the third decimal place.
Units and "interpret" answers
When a question asks for units, they're a point of their own, and you can earn it even with a wrong number. A derivative's units are output units per input unit. An integral of a rate is an amount: rate units times time units. So if R(t) is in liters per minute and t is in minutes, R′(t) is in liters per minute per minute, and ∫₀¹⁰ R(t) dt is in liters.
An interpretation needs three things: what the value measures in the situation, its units, and when (at a specific time, or over a specific interval). Answer exactly what was asked: "the rate is increasing" and "the amount is increasing" are different claims.
Loses the point
The snow is going up by 3.2.
Earns the point
At t = 10 hours, the depth of the snow is increasing at a rate of 3.2 centimeters per hour.
Loses the point
412 liters per minute flow into the tank.
Earns the point
412 liters of water flow into the tank from t = 0 to t = 30 minutes.
An integral of a rate is an amount, not a rate, and it needs its time interval.
The BC topics: where the extra rules are
Everything above holds on the BC exam. The BC-only questions (Q2, Q5 and Q6) add rules of their own, and they're where BC students lose the most points. Here they are, topic by topic.
Series: name the test and check its conditions
A series question is scored like any justification: the conclusion only counts with a reason, and the reason is a named test whose conditions you've checked in writing. Say "the series converges", not "the terms converge".
| To show that… | Say this |
|---|---|
| A series diverges (nth-term test) | lim (n → ∞) aₙ isn't 0, or doesn't exist, so the series diverges. If the limit is 0, this test tells you nothing: terms going to 0 never proves a series converges. |
| A geometric series converges, and its sum | |r| < 1, so it converges, and its sum is a/(1 − r), where a is the actual first term (often not 1). |
| A p-series converges or diverges | Σ 1/nᵖ converges when p > 1 and diverges when p ≤ 1. Give the p. A series like Σ 3/(n + 1) isn't the harmonic series: compare it with the harmonic series instead. |
| Integral test | f is positive, continuous and decreasing for x ≥ N, with f(n) = aₙ, so the series does what the improper integral of f from N to ∞ does. The integral's value isn't the series' sum. |
| Direct comparison | Both series have positive terms. Compare the terms, not the series: 0 < aₙ ≤ bₙ with Σ bₙ convergent shows convergence; aₙ ≥ bₙ > 0 with Σ bₙ divergent shows divergence. Then say what the known series does. |
| Limit comparison | Both series have positive terms, lim aₙ/bₙ is finite and positive, and you say what Σ bₙ does. Never use it on an alternating series. |
| Alternating series test | State all three: the terms alternate in sign, |aₙ| decreases (at least from some point on), and lim |aₙ| = 0. |
| Ratio test | L = lim (n → ∞) |aₙ₊₁/aₙ|. If L < 1 the series converges absolutely; if L > 1 it diverges; if L = 1 the test says nothing, so try another one. |
| Absolute or conditional convergence | Test Σ |aₙ| first. If it converges, the series converges absolutely. If Σ |aₙ| diverges but the series itself converges (often by the alternating series test), it converges conditionally. |
| Interval of convergence | The ratio test for the open interval, centered at the series' center; then each endpoint plugged in and checked with a named test and its conditions; then the interval with the right brackets. |
Intervals of convergence are where the most points slip away. On the 2025 series question, only 14% of BC students earned the point for checking both endpoints correctly and giving the final interval (College Board, 2025 Chief Reader Report).
Loses the point
aₙ₊₁/aₙ = (x − 3)√n/(2√(n + 1)), so (x − 3)/2 < 1, and x < 5.
Earns the point
lim (n → ∞) |aₙ₊₁/aₙ| = lim (n → ∞) |x − 3|√n/(2√(n + 1)) = |x − 3|/2. This is less than 1 when |x − 3| < 2, so the series converges for 1 < x < 5.
With no limit and no absolute values, the inequality loses its left half, and the interval isn't centered at x = 3. Write the limit as n → ∞, not x → ∞.
Loses the point
At x = 1 and x = 5 the terms go to 0, so the series converges at both, and the interval is 1 ≤ x ≤ 5.
Earns the point
At x = 5 the series is Σ 1/√n, a p-series with p = 1/2 ≤ 1, so it diverges. At x = 1 it's Σ (−1)ⁿ/√n: the terms alternate in sign, 1/√n decreases, and 1/√n → 0, so it converges by the alternating series test. The interval of convergence is 1 ≤ x < 5.
The ratio test's limit is 1 at both endpoints, so it can't decide them. Each endpoint needs its own named test with its conditions, and the brackets have to match what you found.
Loses the point
lim (n → ∞) (1/(2n + 5))/(1/n) = 1/2, so it diverges.
Earns the point
Both series have positive terms, and lim (n → ∞) (1/(2n + 5))/(1/n) = lim (n → ∞) n/(2n + 5) = 1/2, which is finite and positive. The harmonic series Σ 1/n diverges, so Σ 1/(2n + 5) diverges by the limit comparison test.
The limit on its own proves nothing. The point needs positive terms, a finite positive limit, and what the comparison series does.
- For a series in powers of x², like Σ x²ⁿ/9ⁿ, the ratio test gives x²/9 < 1, so |x| < 3: the radius is 3, not 9.
- Differentiating or integrating a power series term by term keeps the same radius of convergence. At an x outside the interval, the series doesn't converge, so it can't equal the function there.
- If a question asks whether a series converges at a point, give a clear yes or no, with the reason.
Error bounds: write an inequality chain
BC has two error bounds, and both are scored the same way: an inequality chain that starts with the actual error and ends at or below the target. Never write that the error equals the bound. You don't know the error, only that it's at most the bound.
- Alternating series error bound: if the series alternates and its terms shrink in size to 0, the error is at most the size of the first term you left out (the next nonzero term after the last one you used, not the last one you kept). Say why the bound applies: the series alternates, and its terms decrease in size to 0.
- Lagrange error bound: |f(x) − Pₙ(x)| ≤ (max of |f⁽ⁿ⁺¹⁾(z)| for z between a and x) · |x − a|ⁿ⁺¹/(n + 1)!. The derivative, the power and the factorial all use n + 1, and the word max (or a stated upper bound) has to be there.
- Use ≤ when the bound could equal the target, and < only when it's really smaller. "Bound < 0.01" is wrong when the bound is exactly 0.01.
- If the series alternates, use the alternating bound. Reaching for Lagrange, or for a geometric sum, out of habit is a common way to lose the point.
Loses the point
Error = 1/162, and 1/162 < 0.01.
Earns the point
The series alternates, and the size of its terms, 1/(n! · 3ⁿ), decreases to 0. So |S − 13/18| ≤ 1/162 < 0.01.
The error isn't equal to the first omitted term; it's at most that. And the reason the bound applies (the series alternates, with terms shrinking to 0) is part of the point.
Loses the point
Error = 4(0.3)³/3! = 0.018 < 0.02
Earns the point
|f(0.3) − P₂(0.3)| ≤ (max of |f‴(z)| for 0 ≤ z ≤ 0.3) · (0.3)³/3! ≤ 4(0.3)³/3! = 0.018 < 0.02
The chain has to start with the actual error, |f(0.3) − P₂(0.3)|, and say it's at most the bound.
In 2025, 48% of BC students earned the point for writing the Lagrange bound in the right form, but only 28% earned the point for using it to reach the conclusion (College Board, 2025 Chief Reader Report).
Taylor polynomials
- Center it where you're asked: "about x = 1" means powers of (x − 1).
- A polynomial of degree n stops at the (x − a)ⁿ term. Ending with "+ …" or adding higher-degree terms has cost students the last point.
- The coefficient of (x − a)ᵏ is f⁽ᵏ⁾(a)/k!. You don't have to simplify it.
- Write a derivative's value at its point: f″(0) = 5, never f″(x) = 5 or f″(x) = f″(0).
- Derivatives found from a differential equation need the chain rule every time y appears (a dy/dx factor), and the product rule for products.
- "The first four nonzero terms" means skip any zero terms and write exactly four.
- A general term needs parentheses, the right starting index and (−1)ⁿ in the right place. Build it from a pattern of terms you've checked.
- Differentiate a series term by term with respect to x, not n.
Loses the point
P₂(x) = 3 − 2x + (5/2)x² + …
Earns the point
P₂(x) = 3 − 2(x − 1) + (5/2)(x − 1)²
The first one is centered at 0 instead of 1, and the "+ …" turns a polynomial into a series.
Polar area: square r, and the outer curve goes first
- Area = ½ ∫ r² dθ. Square r, and keep the ½.
- Inside one curve and outside another: ½ ∫ (R² − r²) dθ, where R is the curve farther from the pole on that interval of θ. Check one value of θ in the interval, or look at the graph. It's R² − r², not (R − r)².
- The limits are the angles where the curves meet: solve R(θ) = r(θ), in radians. The setup has to be a definite integral with numbers as its limits.
- Symmetry is fine: twice the area of half the region.
Loses the point
Area = ∫ from −π/3 to π/3 of (4 cos θ − 2)² dθ
Earns the point
Area = ½ ∫ from −π/3 to π/3 of ((4 cos θ)² − 2²) dθ = 7.653
Square each radius, then subtract, and keep the ½. r = 4 cos θ is the outer curve here: at θ = 0 it's 4, farther from the pole than 2.
On the 2025 polar question, only 16% of BC students earned the answer point for the area, the point that checks the ½ and the limits (College Board, 2025 Chief Reader Report).
Motion in the plane: speed, position and distance
- Velocity is the vector ⟨x′(t), y′(t)⟩, and acceleration is ⟨x″(t), y″(t)⟩. Give the one that's asked for, with the components in order.
- Speed is √((x′(t))² + (y′(t))²), never |dy/dx|. To find when the speed equals some value, write that equation and solve it.
- Position: x(b) = x(a) + ∫ₐᵇ x′(t) dt, a definite integral with dt. Working backward from a later time, subtract the integral instead.
- Total distance traveled is the integral of the speed over the interval, the same integral as the arc length of the path. It's not built from the two components' separate integrals.
- The slope of the path is dy/dx = (dy/dt)/(dx/dt) at the right time.
- Moving toward the x-axis means y(t) and y′(t) have opposite signs. Say both signs and the interval.
Loses the point
x(3) = ∫ ln(t + 2) + 4
Earns the point
x(3) = x(1) + ∫₁³ ln(t + 2) dt = 4 + 2.751 = 6.751
The first line has no limits and no dt, so it doesn't say which integral, or which variable.
Loses the point
√((∫₀² x′(t) dt)² + (∫₀² y′(t) dt)²) = 3.971
Earns the point
∫₀² √((ln(t + 2))² + (t² − 3)²) dt = 4.463
The first one is the straight-line distance between the start and the end, not the length of the path the particle took.
"Farthest", "closest" and other polar and motion claims
| To show that… | Say this |
|---|---|
| A point on a polar curve is moving away from the origin | r and dr/dθ have the same sign: r > 0 and dr/dθ > 0, or r < 0 and dr/dθ < 0. "r is increasing" isn't enough when r is negative. |
| A point is the farthest from (or closest to) the origin or an axis | Use the right distance: |r| from the origin, |x| = |r cos θ| from the y-axis. Then a global argument: compare the values at every critical point and both endpoints, or use the sign of the derivative on the whole interval. |
| r has a relative maximum or minimum at θ = c | dr/dθ changes sign at θ = c (or use the second derivative test). Comparing candidate values doesn't earn this one. |
| The rate r changes with time | dr/dt = (dr/dθ)(dθ/dt), by the chain rule. |
| The slope of a polar curve | Use x = r cos θ and y = r sin θ, then dy/dx = (dy/dθ)/(dx/dθ). |
On the 2025 polar question, only 6% of BC students earned the point for a global argument about the farthest point (College Board, 2025 Chief Reader Report). A sign change at one value only shows a relative extremum.
Improper integrals: limit notation on every line
- Rewrite the integral as a limit, such as lim (b → ∞) of ∫₁ᵇ, and keep writing lim on every line until you actually take the limit. Plugging in ∞, or doing arithmetic with it, loses the point.
- An integrand that blows up at an endpoint, or anywhere inside the interval, makes the integral improper too: split it at the bad point and use a limit for each piece.
- If the limit is infinite or doesn't exist, write "the integral diverges".
- With a substitution, change the limits to the new variable correctly.
Loses the point
∫₁^∞ x⁻³ dx = [−1/(2x²)] from 1 to ∞ = 0 + 1/2 = 1/2
Earns the point
∫₁^∞ x⁻³ dx = lim (b → ∞) ∫₁ᵇ x⁻³ dx = lim (b → ∞) (−1/(2b²) + 1/2) = 1/2
Putting ∞ in for x is arithmetic with infinity. The limit has to be written, and kept, until it's taken.
Integration by parts and partial fractions
- Show your u and dv. The tabular method counts too, as long as its columns start with the right functions.
- ∫ u dv = uv − ∫ v du: keep the minus sign, put parentheses around whatever you substitute, and end an indefinite integral with + C.
- Check your v by differentiating it. Coefficients flip easily: the antiderivative of e^(x/2) is 2e^(x/2), not ½e^(x/2).
- Partial fractions: factor the denominator, find the constants, and integrate to logarithms with absolute values, keeping the 1/a in front of ln|ax + b|. If the numerator's degree isn't lower than the denominator's, divide first.
Loses the point
∫ x e^(x/2) dx = ½x e^(x/2) − ¼e^(x/2)
Earns the point
u = x and dv = e^(x/2) dx, so du = dx and v = 2e^(x/2). ∫ x e^(x/2) dx = 2x e^(x/2) − ∫ 2e^(x/2) dx = 2x e^(x/2) − 4e^(x/2) + C
The first one has the wrong v, no u and dv, and no + C.
Euler's method
- Take exactly the number of steps asked, with the step size given, starting from the given point.
- Each step: new y = old y + Δx · (the slope at the point you're at now).
- Write ≈ for the estimates, and show every step.
Loses the point
y(0.2) ≈ 1 + 0.2(2) = 1.4
Earns the point
y(0.1) ≈ 1 + 0.1(2) = 1.2. The slope at (0.1, 1.2) is 0.1 + 2(1.2) = 2.5, so y(0.2) ≈ 1.2 + 0.1(2.5) = 1.45.
Two steps means two steps of size 0.1, and each one uses the slope at the point you're at.
In 2025, 72% of BC students earned the point for the first Euler step, but only 44% earned the final answer (College Board, 2025 Chief Reader Report). A slip in the first step carries through to the end, so check it before you go on.
Logistic models
Logistic models and partial fractions haven't had a free-response part from 2023 to 2026, but they can still come up in multiple choice. Most logistic questions can be answered from the differential equation alone, without solving it:
- In dy/dt = ky(a − y), or dy/dt = ky(1 − y/a), the number a is the carrying capacity: the value y(t) approaches as t → ∞, when y starts above 0.
- If y starts above a, it decreases toward a.
- y grows fastest when y = a/2, halfway to the carrying capacity. Questions usually ask for that value of y, not a time.
Loses the point
The carrying capacity is 0.002, and P grows fastest at t = 300.
Earns the point
lim (t → ∞) P(t) = 600, so the carrying capacity is 600. P is growing fastest when P = 300, half of 600.
The carrying capacity is the value P levels off at, not k. And the question asks for the population, not a time.
When something goes wrong
- Use your earlier answers, even ones you doubt. If part (b) needs your value from part (a), use it: work that's consistent with a wrong earlier answer can still earn the later points.
- Stuck on a part? Move on. Later parts often stand on their own, and you can start fresh from the information in the question.
- Short on time? Write the setup. A correct integral or equation can earn its point even without the value.
- Cross out work you don't want graded with a single line instead of spending time erasing. Crossed-out work isn't scored.
- Commit to one answer. If two different answers to the same part are left standing, the reader won't pick the better one for you.
Before you leave a question
- Did I answer what was asked, in the form it asked for: a value, an x-value, an interval, a sentence?
- Is every number backed by work, and every calculator result by its setup?
- Does every reason name its function, and check the conditions of any theorem I used?
- Three decimal places? Units, if the question asked for them?
- Is there a stray extra answer, or an = that isn't true?
And on the BC-only questions, also check:
- Series: lim (n → ∞) written, absolute values in the ratio, both endpoints tested with named tests, and the brackets right?
- Error bound: a chain that starts at the actual error and ends at the target, with no "error ="?
- Taylor polynomial: the right center, the right degree, and no "+ …"?
- Polar area: the ½, each r squared, the outer curve first, and θ-limits where the curves meet?
- Motion: speed from both components, and each position as a definite integral with dt?
- Improper integral: lim on every line until it's taken?
- Euler: the right number of steps, each slope taken at the current point?
Calculator skills
You need a graphing calculator in two parts only: multiple-choice Part B and free-response Part A. In those parts Bluebook has a built-in Desmos graphing calculator, and you can also bring up to two approved handheld graphing calculators. The rest of the exam is calculator-free.
Every approved calculator can do four things, and these four are the only calculator results you can use without showing the steps. Practice them until they're automatic:
- Graph a function in a window you choose.
- Solve an equation numerically, which includes finding where a function equals zero.
- Find the value of a derivative at a point.
- Find the value of a definite integral.
What to write when you use it
On a free-response question, write the setup in ordinary math notation (the equation you solved, the derivative you evaluated, or the integral you computed), then the result. Calculator commands aren't math notation and don't earn the setup point. Anything beyond the four skills, such as a built-in maximum finder, a regression, a table search or a program, doesn't count as a reason: show the calculus instead.
Loses the point
fnInt(ln(X), X, 1, 3) = 1.296
Earns the point
∫₁³ ln x dx = 1.296
Loses the point
intersect: x = 0.739
Earns the point
cos x = x when x = 0.739
Loses the point
Calculator max: f has a relative maximum at x = 1.571.
Earns the point
f′(x) = 0 at x = 1.571, and f′(x) changes from positive to negative there, so f has a relative maximum at x = 1.571.
A maximum finder isn't one of the four skills, so its result isn't a reason.
Desmos in Bluebook (multiple choice, Part B)
- Try the built-in Desmos in a Bluebook test preview before exam day. The Desmos website and app aren't allowed during the exam, but you can practice the same skills there.
- Define a function once, like f(x) = …, and reuse it: type f'(2) for the derivative's value at 2, and type int to start a definite integral.
- To solve an equation, graph each side as its own curve and tap the point where they cross. For zeros, tap where the graph crosses the x-axis.
- Set the window with the zoom buttons or the graph settings, then zoom out once to make sure you haven't missed a crossing.
- A point's label can show only a few decimal places. If two answer choices are close, zoom in on the point to see more digits.
- Desmos works in radians unless its settings are switched to degrees. Leave it in radians.
- Not every Part B question needs the calculator. If you can see the answer, take it and save the time.
Your handheld (free response, Part A)
In free-response Part A you can use the built-in Desmos or a handheld calculator. If you use a handheld:
- Set radian mode before the exam, and check it again when Part A starts.
- Enter each given function once, in the calculator's function list, and evaluate from there instead of retyping a long formula.
- Store intermediate results in a memory variable (A, B, and so on) instead of copying a rounded number, and use the stored value in the next step.
- Learn where your model keeps each of the four tools, and practice until you reach them without hunting through menus.
- Write the setup in the booklet first, then compute. If the calculator gives something odd, the setup shows what you meant.
- Bring fresh batteries or a full charge. A second approved calculator is allowed as a backup.
What BC adds
The calculator question on the BC side, Q2, has been about motion in the plane or a polar curve, and both lean on the calculator. Practice these on the calculator you'll bring and in Bluebook's Desmos:
- Graph a parametric curve. In Desmos, type the point (x(t), y(t)) and set the range of t; on a handheld, switch to parametric mode. The graph shows the path and which way it goes.
- Graph a polar curve. In Desmos, type r = …; on a handheld, switch to polar mode. Set the range of θ to the one in the question.
- Find where two polar curves meet by solving R(θ) = r(θ) for θ, in radians: those angles are your area limits. Curves can also meet at the pole, where r = 0, at different angles, so check the graph.
- Let numeric integrals do the heavy lifting: polar area, total distance, arc length, and a position found from a velocity. Enter each integrand as an ordinary function of one variable. The College Board's 2025 Chief Reader Report suggests evaluating a polar area integral in function (rectangular) mode, not polar mode.
- Use the numeric derivative for a velocity, an acceleration or a value of dr/dθ at one point.
- To find when the speed equals some value, solve the equation √((x′(t))² + (y′(t))²) = value with the solver, or by graphing both sides, and take the time that fits the question. TRACE isn't accurate enough.
- Store each value you'll reuse, like an angle where two curves meet, and use the stored value in the next step.
- Radian mode matters on every BC calculator question. In the 2023 BC scoring guidelines, a response worked in degree mode lost the first point it would otherwise have earned, and polar angles in degrees give the wrong area limits.
Loses the point
TRACE: the speed is about 2 at t ≈ 1.17.
Earns the point
√((ln(t + 2))² + (t² − 3)²) = 2 at t = 1.169
Write the equation you solved. TRACE isn't one of the four calculator skills, and it isn't accurate to three decimal places.
Loses the point
4 cos θ = 2 at θ = 60
Earns the point
4 cos θ = 2 at θ = −π/3 and θ = π/3 (about ±1.047)
Degrees give the wrong limits, and with them the wrong area.
Which calculators are allowed
A handheld has to be on the College Board's list of approved graphing calculators for AP Calculus. The list marks the models that have the four required capabilities. Examples include the TI-84 Plus CE, TI-89, TI-Nspire CX and CX CAS, HP Prime and NumWorks; check your exact model on the official list.
These can't be used as your calculator: phones, watches, laptops and tablets, devices with a QWERTY keyboard or pen or stylus input, anything with a wireless, Bluetooth, cellular or internet connection, and touch-screen models that aren't on the list. You don't have to clear your calculator's memory, but you can't share a calculator with anyone.
A calculator with a computer algebra system (CAS) is allowed if its model is approved, but you don't need one: nothing in this course needs it.
What your score means
AP scores run from 1 to 5. The College Board describes each score like this:
| Score | What it means |
|---|---|
| 5 | Extremely well qualified |
| 4 | Well qualified |
| 3 | Qualified |
| 2 | Possibly qualified |
| 1 | No recommendation |
Each college sets its own policy. Many give credit or placement for a 3 or higher, and some require a 4 or 5, so check the policy of every college you're considering.
The AB subscore
Your BC score report also shows a Calculus AB subscore, from 1 to 5. It's based on the part of the BC exam that tests AB topics, about 60% of the exam. So the AB topics count twice: toward your BC score and toward the subscore.
The College Board asks colleges to treat the AB subscore the way they treat an AP Calculus AB score, because the shared topics are tested at the same level in both courses. Each college still sets its own policy, so look up both: what it gives for your BC score, and what it gives for the AB subscore. Exactly which questions count toward the subscore isn't published.
Recent results
Swipe the table sideways to see every column.
| Year | 5 | 4 | 3 | 2 | 1 | 3 or higher | Mean score |
|---|---|---|---|---|---|---|---|
| 2026 | 46% | 22% | 14% | 14% | 4% | 82% | Not listed yet |
| 2025 | 44.0% | 21.9% | 12.8% | 15.2% | 6.2% | 78.6% | 3.82 |
| 2024 | 47.7% | 21.1% | 12.1% | 13.9% | 5.2% | 80.9% | 3.92 |
| 2023 | 43.5% | 15.9% | 19.0% | 15.2% | 6.3% | 78.5% | 3.75 |
| 2022 | 41.2% | 15.6% | 20.1% | 16.4% | 6.8% | 76.9% | 3.68 |
| 2021 | 38.3% | 16.5% | 20.4% | 18.2% | 6.6% | 75.2% | 3.62 |
Source: the College Board's AP Students website. The 2026 figures are listed in whole percents so far and may be updated. In 2025, 160,436 students took AP Calculus BC.
Swipe the table sideways to see every column.
| Year | 5 | 4 | 3 | 2 | 1 | 3 or higher | Mean subscore |
|---|---|---|---|---|---|---|---|
| 2025 | 47.7% | 30.5% | 9.8% | 9.0% | 3.0% | 88.0% | 4.11 |
| 2024 | 50.1% | 29.7% | 8.4% | 9.2% | 2.6% | 88.2% | 4.16 |
| 2023 | 46.4% | 19.6% | 18.8% | 8.6% | 6.6% | 84.8% | 3.91 |
Source: the College Board's AB subscore distributions for each year. The 2026 subscore figures hadn't been posted when we checked in late October 2026. No subscores were reported in 2020, when the exam was shortened.
How the score is worked out
Your multiple-choice and free-response results are combined into one composite score, each section worth half. The College Board then decides how many composite points each AP score takes. Those cut points are set every year, they change from year to year, and the College Board doesn't publish them for recent exams. Old released exams printed conversion tables, but those exams had a different design.
The traditional estimate gave each multiple-choice question 1.2 points, so 45 questions made 54. With 42 questions, keeping the same 50/50 split means multiplying the number you got right by 54/42 (about 1.286) and adding your free-response points, for 108 points in all. It's an estimate, not an official formula.
That holds for BC too, and the College Board doesn't publish how the AB subscore is worked out either. Online score calculators make their own guesses. Here are three of them, out of the 108-point composite above:
Swipe the table sideways to see every column.
| Score | Test Ninjas and Num8ers (2026) | AdmissionSight |
|---|---|---|
| 5 | About 68 or more (about 63%) | About 62 or more |
| 4 | About 56 or more (about 52%) | About 47 or more |
| 3 | About 42 to 44 or more (about 39% to 41%) | About 34 or more |
| 2 | About 35 or more | Not given |
Source: each site's AP Calculus BC score calculator, checked in October 2026. These are third-party guesses, not College Board numbers, and they disagree by several points.
Study plan
The plan below counts down to the exam on Monday, May 10, 2027. It shares out the weeks you have left among the units you still need to study (all 10 of them to start with), in course order, and gives more time to the units that carry more weight on the exam. The last weeks are kept for review and a full practice exam.
When you're signed in, the plan skips the units you've already finished and gives their time to the rest. Treat it as a guide: if you're ahead, start the next unit; if you fall behind, keep going in order and protect the review weeks.
The 10 units, and what BC adds to each
| Unit (share of BC multiple choice) | What BC adds |
|---|---|
| Unit 1: Limits and Continuity (5–10%) | The same topics as AB |
| Unit 2: Differentiation: Definition and Fundamental Properties (5–10%) | The same topics as AB |
| Unit 3: Differentiation: Composite, Implicit, and Inverse Functions (5–10%) | The same topics as AB |
| Unit 4: Contextual Applications of Differentiation (5–10%) | The same topics as AB |
| Unit 5: Analytical Applications of Differentiation (10–15%) | The same topics as AB |
| Unit 6: Integration and Accumulation of Change (15–20%) | Integration by parts, partial fractions, improper integrals |
| Unit 7: Differential Equations (5–10%) | Euler's method, logistic models |
| Unit 8: Applications of Integration (5–10%) | Arc length |
| Unit 9: Parametric Equations, Polar Coordinates, and Vector-Valued Functions (10–15%) | All of it: parametric and vector motion, polar curves and polar area |
| Unit 10: Infinite Sequences and Series (15–20%) | All of it: series tests, Taylor polynomials and series, error bounds |
The shares are for the multiple-choice section, from the College Board's course description; free-response questions draw on the whole course. Unit 10 is the longest: the College Board's suggested pacing gives it about 17 to 18 of roughly 115 class periods.
Where to spend extra time
Swipe the table sideways to see every column.
| Year | Q1 | Q2 (BC only) | Q3 | Q4 | Q5 (BC only) | Q6 (BC only) |
|---|---|---|---|---|---|---|
| 2025 | 5.22 | 3.09 | 6.27 | 5.46 | 5.21 | 4.32 |
| 2024 | 6.45 | 5.56 | 5.57 | 5.84 | 5.92 | 3.38 |
| 2023 | 5.26 | 5.20 | 3.74 | 4.03 | 4.71 | 4.21 |
The two lowest means from 2023 to 2025 were both BC only: the 2025 polar question (3.09) and the 2024 series question (3.38). So plan extra free-response practice for Units 9 and 10, and once you've finished Unit 10, keep doing a series question every week or two until May.
- Study a little on most days rather than a lot on one day.
- Finish each unit with its review and timed unit test, then go back to the topics still marked Practicing on the course page.
- Write your free-response practice by hand, on paper, the way you'll write it in May.
- Practice the four calculator skills on the calculator you'll bring, and try the built-in Desmos in a Bluebook test preview.
- In the review weeks, take the full practice exam in one sitting with real timing, then redo every question you missed.
- Took AB already? Use the Unit 1 to 5 reviews and timed tests to check what you remember, and move on quickly where you score well.
- There's no formula sheet, so make flash cards for the formulas in How the exam works, and say the series tests' conditions out loud until they're automatic.
- Practice graphing parametric and polar curves on the calculator you'll bring, and in the Bluebook Desmos.
- Write series justifications in full every time, even in practice: the limit, the test's name, its conditions, and both endpoints.
Your plan: 26 weeks to the exam
Starting this week and ending the day before the exam (May 10).
Sign in and the plan skips what you've already finished.
Week 1
Nov 11 – Nov 17
Week 2
Nov 18 – Nov 24
Week 3
Nov 25 – Dec 1
Week 4
Dec 2 – Dec 8
Week 5
Dec 9 – Dec 15
Weeks 6–7
Dec 16 – Dec 29
Weeks 8–9
Dec 30 – Jan 12
Week 10
Jan 13 – Jan 19
Weeks 11–13
Jan 20 – Feb 9
Week 14
Feb 10 – Feb 16
Week 15
Feb 17 – Feb 23
Week 16
Feb 24 – Mar 2
Week 17
Mar 3 – Mar 9
Weeks 18–20
Mar 10 – Mar 30
Weeks 21–24
Mar 31 – Apr 27
Week 25
Apr 28 – May 4
Review
Redo the unit tests, then practice the topics still marked Practicing on the course page.
Week 26
May 5 – May 9
Full practice exam
Take it in one sitting with real timing, then redo every question you missed. Rest the day before the exam.
Each unit: its lessons, then the unit review and the timed unit test. Time per unit follows how much of the exam the unit is worth. Ahead of schedule? Great: start the next unit.
Unit 1 is free with an account.
AP® is a trademark registered by the College Board, which is not affiliated with, and does not endorse, this course.