Written-response paper2026 · Main version · Exam-day written responses
- Document
- Complete original PDF
- Pages
- 4
- File size
- 0.17 MB
AP CSP · PROGRAMS · COMPUTATIONAL THINKING
Understand your code.
Explain the decisions behind it.
Browse AP Computer Science Principles written-response papers and Create performance-task resources. Find a complete released paper, its matching scoring guide or a historical written sample. Read each original document on screen for practice.

Marked a full practice exam? Use the AP Computer Science Principles score calculator to turn your marks into an estimated 1–5 score.
FULL PAPERS AND CREATE RESOURCES
Start with the released written-response papers. Choose All documents for historical samples, scoring commentary and reports, then narrow your selection.
58 complete documents available · choose written-response papers or supporting resources
No documents match. Select All documents, clear the version filter or choose another year.
PREPARING FOR MAY 2027
AP CSP assesses a broad understanding of computing as well as the ability to develop and explain a program. The current end-of-course exam lasts three hours and is fully digital in Bluebook. The Create performance task is developed during the course; its assessment is connected to the exam-day written responses.
| Component | Time | Contribution to AP score |
|---|---|---|
| Section I: multiple choice | 70 questions · 120 minutes | 70% |
| Create program, video and project reference | At least 9 hours of class time | 30% combined with written responses |
| Section II: written responses | 2 questions containing 4 prompts · 60 minutes |
The multiple-choice section contains 57 single-select questions, five single-select questions linked to a reading passage about a computing innovation, and eight multiple-select questions requiring two answers. A released written-response PDF does not include these 70 questions, so a one-hour attempt with it is not a complete mock exam.
The written prompts ask you to reason about your own Create program using your Personalized Project Reference (PPR). Practise with code you understand and the reference you prepared for a practice project. A generic explanation borrowed from a sample cannot establish how your own algorithm behaves.
The current task has three submissions: program code, a video demonstrating the program running, and the PPR. Each must be submitted as final in the AP Digital Portfolio. The reference contains selected screenshots showing the required procedure and call, plus list storage and use; it is not the same document as the full program-code PDF.
Use the official student handouts for exact requirements. Code screenshots for the PPR must remain readable and must not contain comments or course content. The full code submission has different requirements, including appropriate acknowledgments. The video and PPR are completed individually; permitted collaboration on program development does not extend to those two components.
For practice, check that you can locate the relevant code and explain it without a running editor. Distinguish a procedure definition from a call, a parameter from the argument supplied to it, and the collection being stored from the code that uses it. If your explanation depends on hidden context, revise your practice reference and review the current directions.
The 2018–2023 sample PDFs in this library show historical Create written submissions. Current exam-day responses replaced that earlier submission arrangement from the 2024 exam. The older work can help you examine clarity, testing and abstraction, but its response numbering and word limits should not be treated as current exam instructions.
Read those samples with the commentary for the same year. Several commentary covers say that student samples are provided separately; a commentary PDF is therefore not evidence that this collection includes every sample or its video. Your practice should connect what you can actually see in a document to the rubric that evaluated it.
College Board has announced an AP CSP redesign for the 2027–28 school year. Its published course-change overview lists the current fall 2023 course description for 2026–27, with a redesigned description planned for spring 2027. The May 2027 structure above applies to the current course, not the later redesign.
The announcement describes additional AI content and an AI Create Project for the later course. It does not make these historical PDFs a complete guide to the redesigned assessment. Check the official change overview and your teacher’s current course materials before applying an old timing, rubric or project requirement to a later exam.
PAPERS, SAMPLES AND SCORING
The library holds five written-response papers and 53 supporting PDFs. Both released sets are available for 2024 and 2025; 2026 has one paper. The resources span nine years, but only 2024–2026 have exam-day prompt papers in this collection.
| Year | Written-response papers | Scoring guides | Separate written samples | Other review PDFs |
|---|---|---|---|---|
| 2026 | 1 | 0 | 0 | 0 |
| 2025 | 2 | 2 | 0 | 5 |
| 2024 | 2 | 2 | 0 | 5 |
| 2023 | 0 | 1 | 10 | 3 |
| 2022 | 0 | 1 | 0 | 3 |
| 2021 | 0 | 1 | 9 | 3 |
| 2020 | 0 | 0 | 0 | 1 |
| 2019 | 0 | 1 | 0 | 3 |
| 2018 | 0 | 0 | 3 | 0 |
Choose the same year and Set 1 or Set 2 for the paper, scoring guidelines, reader report and scoring statistics. The sets have different prompts; mixing them can lead you to mark an answer against a task you did not attempt. Select Shared resources to find yearly distributions and historical documents without a set label.
The 2024 and 2025 collections each contain two papers, two guides, two reader reports, two statistics files and one distribution. Neither collection includes the separately released sample-response packages. For 2026, no scoring guide is supplied here, so choose an earlier paired paper if you want to review an attempt against its own guide.
2018 contains Create written samples A, B and C only. There is no accompanying guide, commentary or video in that year’s collection. The 2020 folder contains one reader report. The 2021 samples run A–F and H–J: sample G is absent. The 2022 resources include a guide and commentary, but no separate student samples.
All ten separate samples A–J are available for 2023. The four commentary documents cover 2019, 2021, 2022 and 2023. They are preserved as complete publisher PDFs, with any included scoring material, but are labelled as commentary rather than counted again as separate exam papers.
| Document type | PDFs | How to use it |
|---|---|---|
| Scoring guidelines | 8 | Apply the rubric for the matching year and version. |
| Historical Create written samples | 22 | Study how students explained their programs under earlier directions. |
| Create scoring commentary | 4 | Read the explanation of scores alongside the separate samples, where supplied. |
| Reader reports | 9 | Review recurring difficulties in program explanations and assessment. |
| Scoring statistics | 4 | Read the reported results for the relevant exam set. |
| Score distributions | 6 | See the reported spread of final AP scores for that year. |
No sample videos, full project-code packages or complete multiple-choice papers are included. A written sample may refer to a video or program that is not present in its PDF. Treat that as a limit on your review rather than imagining the missing evidence. The official archive at the end can help you find further released materials.
FROM PROGRAM TO EXPLANATION
Begin with a small program developed for practice that uses a meaningful collection and a procedure. Run it with several inputs and record what it does. Then explain the same behavior from the code alone. This separates an observation about one run from a claim about what the algorithm will do in general.
Decide what the program is intended to achieve for its user. Its purpose is not simply the sequence of statements it executes. Describe how an input affects the result and what output the user receives. When discussing a particular code segment, connect it to the larger purpose rather than listing syntax without context.
Open a 2024 or 2025 paper if you want an available matching guide. Keep the guide closed while attempting the entire written section with your practice project reference. Use a 60-minute allowance and type your explanations. Reserve some time to check that each response addresses the wording in the paper and refers to the intended code.
When a prompt asks about a modification, reason through the resulting behavior rather than only describing the original program. When it asks for an explanation, state the relationship that makes your claim true. A program can run successfully while an explanation remains incomplete or inconsistent with the code.
Follow statements in execution order. Keep a small trace table of changing values, the condition being tested and any output. For iteration, check the initial state, each update and the condition that ends repetition. Evaluate a conditional when execution reaches it; do not assume that a variable still has its earlier value.
A procedure may be called repeatedly with different arguments. Trace one call at a time, noting the parameter values and the returned value or side effect. Use different test inputs to explore which branch runs. Explain why a change affects the result, including the steps where that effect first becomes visible.
For a list, identify the related data it represents and how the program accesses or transforms those values. Explain the practical difference between using that collection and handling the same data with separate variables. Refer to the actual operations your program performs, rather than claiming that every list automatically makes every program simpler.
For procedural abstraction, describe what task the procedure packages and how its parameters allow reuse. Consider the code that would otherwise be repeated at call sites. An explanation becomes more useful when it identifies a concrete change that can be made in one place and shows which calls benefit from that change.
Choose input values deliberately. State the expected output before running the program, then compare it with the actual result. Include normal cases and values near a relevant boundary. If a selection condition separates two behaviors, choose inputs that exercise each behavior instead of repeating the same path.
Separate an error from its symptom. An unexpected output tells you that something went wrong, but it does not identify the statement responsible. Trace back to the first value or decision that differs from the intended behavior. Explain how a proposed correction addresses that cause and which test would show that the correction worked.
Attempt your own explanation before reading a sample. Then compare the specificity of its reasoning with yours. Look for how the writer connects a code segment to purpose, explains a collection or justifies a test. These habits can remain valuable even when the assessment directions have changed.
Read the matching historical commentary to understand why particular evidence earned or lost credit. Where a sample refers to an unavailable video, acknowledge that you cannot independently check its demonstration. Use the commentary as a report of the original scoring, not as permission to infer unseen code or functionality.
Written-response practice alone does not cover the full computing course. Include conceptual questions, code tracing and evaluation of computing innovations in your revision. Use teacher-assigned AP Classroom resources and the course description for authorized practice, and use the official Bluebook preview to learn the digital interface.
| Big idea | MCQ weighting | Revision focus |
|---|---|---|
| Creative Development | 10%–13% | Purpose, design choices, development and collaboration. |
| Data | 17%–22% | Representation, transformation and drawing conclusions from data. |
| Algorithms and Programming | 30%–35% | Control flow, procedures, collections and algorithm behavior. |
| Computer Systems and Networks | 11%–15% | How connected systems communicate and continue to function. |
| Impact of Computing | 21%–26% | Benefits, risks and effects of computing innovations. |
These ranges describe coverage within Section I, not extra percentages added to the final AP score. Balance time across the course according to both the assessment and your own weaknesses. Familiarity with one programming project does not replace understanding of data, systems and the wider effects of computing.
After marking, classify what caused a difficulty: misunderstanding the prompt, tracing code incorrectly, using the wrong segment or omitting the reasoning. Rewrite the explanation that failed and test it on another program or input. Improving a specific weakness is more informative than repeatedly reading an answer you already remember.
Keep your practice marks separate from a final AP prediction. A written paper does not provide the multiple-choice result or all the evidence used to assess Create. Reader statistics describe a group’s performance, while a score distribution reports final grades; neither guarantees a conversion for your practice attempt.
For a structured revision routine, use our guides to learning from exam mistakes and building a study week. Browse other collections in Resources, or explore AP tutoring and coding support for practice in computational thinking and program explanation.
A LITTLE MORE CLARITY
Choose the correct documents and use historical Create resources appropriately.
Tell us about your needsThis library contains five complete written-response papers from 2024–2026 and 53 supporting PDFs, with resources spanning 2018–2026. Select a year, version and document type, then read the original PDF.
No. The five released papers contain the written-response section only. They do not contain the 70-question multiple-choice section or provide a complete three-hour mock exam. Use authorized AP Classroom practice and the course description alongside them.
No. Each card represents a complete original document. Read PDF starts at page 1, and the popup lets you navigate through that document. Samples A–J are publisher sample labels, not extracted questions.
There are no exam-day written-response papers before 2024 in this collection. The 2018 files are samples A–C only; 2020 has one reader report. The 2021 separate samples omit G, and 2022 has commentary but no separate sample files. No sample PDFs or videos are supplied for 2024–2026; 2026 has its paper only, without a matching scoring guide.
The 2024 and 2025 written-response papers have two released exam versions. Match the year and set when selecting scoring guidelines, reader reports and statistics. A yearly score distribution is shared across versions; it is not a separate Set 1 or Set 2 answer key.
The current end-of-course exam lasts three hours and is fully digital in Bluebook. Section I has 70 multiple-choice questions in 120 minutes and contributes 70%. The Create performance task and 60-minute written-response section together contribute 30%. Two written-response questions contain four prompts and relate to your own project reference.
The PPR is a student-authored reference containing selected screenshots from the program developed for the Create performance task. It shows the required procedure and its call, and list storage and use. Submit it as final through the AP Digital Portfolio to make it available on exam day; follow the current handout rules for readable code without comments or course notes.
Use them to study programming explanations and historical scoring, with their matching commentary. The 2018–2023 written samples use earlier through-course response arrangements. They are not current PPR templates or exam-day answer keys, and the collection does not include their videos or complete project submissions.
The redesign is announced for the 2027–28 school year. The May 2027 format above belongs to 2026–27. Check the official course-change reference at the end for later updates; historical papers do not establish the redesigned exam structure or weighting.
It cannot determine the final score on its own. Multiple-choice performance and the Create assessment also matter. Use the matching rubric to identify improvements; statistics and score distributions do not supply a guaranteed raw-mark-to-AP-score conversion.