How to Study for the AP Latin Exam

Half of AP Latin rests on a published list of Vergil and Pliny passages you can genuinely finish reading before May. The other half is Latin you have never seen, which no amount of memorising reaches. Those two halves need different hours, and one graded question needs a third skill again.

All of the numbers here are taken from the course and exam description effective Fall 2025, together with College Board’s exam page for the course and its exam calendar. That effective date carries more weight here than in most AP subjects. The required author list was revised for Fall 2025 and first examined in May 2026, so any guide naming a required prose author other than Pliny the Younger is describing a retired exam. Check the date on anything you download before you build a reading plan on it.

What exactly is on the AP Latin reading list?

A finite, published list, which is the unusual part of this course. The required Latin is:

• Pliny the Younger, Letters. Book 6, letters 4, 7, 16 and 20; Book 7, letter 27; Book 10, letters 5, 6, 7, 37 and 90.
• Vergil, Aeneid. Selections from Books 1, 2, 4, 6, 7, 11 and 12.
• Teacher’s Choice. Approximately 1,650 words of prose and approximately 1,350 words of poetry, chosen by your own teacher.

Ten letters and a set of Aeneid selections. College Board lists no required readings in English at all, so every line of that list is Latin you are expected to meet as Latin. The six units of the course are built around those texts, taking Pliny through Vesuvius, ghosts and apparitions, and the letters to Trajan and Calpurnia, alongside the Aeneid selections and the Course Project.

Now notice what that makes possible. In a content-heavy AP you are permanently sampling, because the material is larger than the year you have. Here the syllabus half is a fixed quantity, and reading all of it closely is a goal you can actually complete and then verify. Twenty-six percent of the whole exam is multiple choice on passages from that list, and a passage you have worked through four separate times cannot surprise you in May. Coverage stops being the hard problem here. Decay takes its place, because material fades on a predictable curve, so a letter finished in October needs revisiting long before it needs re-reading from scratch.

How much of the exam is Latin you have never seen?

Just under a quarter of it, and all of that sits in Section I. The course and exam description states the paper plainly: “The exam is 3 hours long and includes 52 multiple-choice questions and 5 free-response questions.” Section I is those 52 questions in 65 minutes for 50 percent of your score, and it divides like this.

Add the sight rows and you get 26 questions worth 24 percent of the exam, on Latin selected precisely so that you have not read it. Add the syllabus rows and you get the other 26 questions, worth 26 percent, on the list you can finish. Section I is cut almost exactly down the middle, and the two halves answer to completely different preparation.

A sight passage is reached by having read a large volume of Latin at a level you can mostly follow, often enough that ordinary constructions stop needing to be decoded one word at a time. That is a fluency problem, and fluency comes from frequent contact spread across months. The daily-reps case in how to study a foreign language transfers here almost unchanged, with the single difference that nobody is going to speak this one back to you.

How do the five free-response questions divide up?

Into one short-answer question, one translation and three short essays. Section II runs 115 minutes and carries the remaining 50 percent of the score, of which the five questions account for 48 percent between them. The last 2 percent comes from the Course Project, whose two in-class checkpoint tasks are done away from exam day.

Two details in that table are regularly misreported. There is no question on this paper called an analytical essay, whatever an older revision site calls it; there are three short essays and one translation. And the suggested timings total 115 minutes, which is the whole of Section II with nothing left over. College Board publishes no separate reading period, so plan on the time you spend reading each passage coming out of the minutes attached to its own question.

What does a 15-segment translation reward?

Accuracy of rendering, checked in fifteen small pieces. Question 2 asks for a literal translation and is scored across 15 segments, each a short stretch of Latin that either arrives in English intact or does not. A fluent paraphrase that quietly loses a tense, a case or a number loses that segment, and no amount of insight elsewhere in the passage buys it back.

That makes it a genuinely separate skill from the three essays. The essays ask what a passage is doing; this question asks whether you read the Latin. Students who are comfortable writing about Vergil often lose segments to exactly the mechanical details they stopped consciously checking a year ago. The drill follows straight from the scoring.

Step four is the one people skip, and it is where the spacing lives. Checking your own translation produces a list of specific failures, and that list is worth more days later than it is in the moment you write it. The case for retrieving material after a gap is what turns a marked-up passage into vocabulary you still hold in May.

Can the published score distributions tell you anything?

Less than usual, and the revision is why. Every distribution College Board has published measures the retired version of this exam. The current framework took effect in Fall 2025 and was first examined in May 2026, and no distribution for that sitting has been released. Read the numbers below as background about the cohort. They describe a paper that has since been replaced.

That comes to 58.6 percent scoring 3 or higher, with a mean of 2.84 across 4,339 students. The year before, 56.5 percent of 4,264 students reached 3 or higher. Two things survive the caveat about the framework. The first is scale: roughly 4,300 candidates a year makes AP Latin one of the smallest AP subjects, small enough that your school may have no recent cohort to measure yourself against and that third-party practice material is thin on the ground. The second is where the mass sits. A 3 is the single most common outcome and a 2 is the next, so most of the cohort lands close to either side of the same line.

Thin third-party material has a practical consequence. Your own timed attempt at a full paper becomes the measurement, and it has to be done early enough that a bad result is still repairable. Sitting an unseen paper under real conditions is the only honest reading you are going to get on the sight half.

What can you carry into the room?

No dictionary. College Board lists among the items you must not bring “correction fluid, dictionaries, highlighters, or notes”, and there is no calculator on this exam. The paper is fully digital and taken in Bluebook.

What stands in for the banned dictionary is printed on the paper itself, under a rule the course and exam description states outright. Appendix 2 is the Full Required Latin Vocabulary List, and the sentence governing it reads: “All Latin words outside this list will be glossed on the AP Exam, except (1) proper names found in the essential knowledge statements, when they occur in the nominative, and (2) on the Project Passage Short Essay free-response questions.”

Read that second exception slowly, because it is the one that catches people. The Project Passage Short Essays are Questions 4 and 5, worth 9 percent each, so on 18 percent of the paper nothing is glossed at all. Everywhere else, across all 52 multiple-choice questions and the first three free-response questions, a word outside the required list arrives with its meaning attached, with the one standing exception below.

That one sentence sets your practice conditions, and it sets two different ones. For Section I and for Questions 1 to 3, work at exactly the level of help the exam gives you: the required vocabulary list held in your own memory, every other word glossed where it stands. Translating the syllabus at home with a dictionary open all evening builds a habit the room takes away from you, and the sight passages are where you feel it first.

For the project passages, take the support away. Read them unaided, with nothing beside you, because unaided is how they reach you in May. That is also the most repairable weakness on this list, since you know which passages they are and you have the whole year to sit with them. The first exception is smaller and still worth an evening: a proper name drawn from the essential knowledge statements goes unglossed anywhere on the paper when it appears in the nominative, so the people, places and gods inside your required reading have to be recognisable on sight in that form.

When does the closed half need to be finished?

Well before exam day, which is Friday, May 14, 2027, in Session 2, the afternoon slot. The 2027 window runs May 3 to 7 and then May 10 to 14, so AP Latin lands on the final day of the second week, about as late as an AP exam can be placed.

Plan around that slot in both directions. It hands you a week of clear air after most of your other papers, which is genuinely valuable for a subject that rewards steady contact. It also means a fortnight of other exams sits between your last unhurried revision block and this one, so whatever you leave half-learned in late April gets no attention for two weeks. Aim to have read every passage on the required list by the start of that window, and keep the final fortnight for translation practice, sight reading and the essays.

Where does a question generator actually help here?

On the memorisable half, and nowhere else. Vocabulary, forms and the content of the syllabus passages are all things a quiz over your own notes can check, and they are exactly what a second read-through of a letter leaves untested. Sight-reading fluency is not on that list, and no question generator translates for you.

Upload a document to GeniusPal and it writes questions from what is inside it. Feed it your notes on the Vesuvius letters, or your own glossary for a block of the Aeneid, and a page you had merely reread becomes a score you can argue with. For a closed reading list that is a good fit, because the thing you have to know is finite and can be checked repeatedly.

The limits are worth stating exactly. A free account carries two study-set generations for the lifetime of the account, 10 questions per set, every question multiple choice, two full runs of each set, and one exam-style analysis, also for the account lifetime. Student is $14.99 a month and raises that to 100 generations, up to 30 questions per set, ten exam-style analyses a month, and the option of taking the same set as flashcards or as recall. Quiz, flashcard and recall are the only three study modes there are; written short answers are a format inside the quiz on paid plans. Uploads go up to 10 MB on every plan.

Two of its mechanics are free everywhere and fit a year-long vocabulary load well. The daily review carries 20 questions drawn from what you got wrong or have not tried yet. Redrilling wrong answers is free on every plan too. Between them they give you the cheap repeated checking that a list of forms and glosses needs between October and May.

Exam-style analysis runs in three steps and is never a prediction. A past paper is read on its own into a saved style. You must review that style before it shapes anything. You then pick it on a later upload, so questions generated from your notes follow that shape. The effect shows up most in written answers, which need a paid plan, so on a free account the wording of a multiple-choice set moves much less. It tells you nothing about what will appear on your exam.

So use it for the layer underneath: the vocabulary, the forms, and what happens in which letter. Then spend the hours it saves on the two things it cannot do, which are translating segment by segment with your own hand and reading Latin you have never seen before.