Free Practice Questions for Claude Certified Developer - Foundations

Exam-style questions and explanations by OpenExamPrep.

✓ No registration✓ No credit card

164+ Questions

100% Free

Loading practice questions...

Same family resources

Explore More Anthropic Claude Certifications

Continue into nearby exams from the same family. Each card keeps practice questions, study guides, flashcards, videos, and articles in one place.

More From This Family

Videos and articles for deeper review.

VideoCCA-F Practice Questions 2026: Current CCAR-F Blueprint, Scenarios, and Answer Strategy VideoCCAR-F Exam Guide 2026: Claude Certified Architect – Foundations VideoCCAR-F vs CCA-F: Registration and Pearson VUE Guide for Claude Architect Foundations (2026)VideoWhich Claude Certification Should You Take in 2026? CCAO-F vs CCDV-F vs CCAR-F vs CCAR-P ArticleCCAO-F Exam Guide 2026: Claude Certified Associate Domains, Study Plan, and Free Practice18 min read ArticleWhich Claude Certification Should You Take in 2026? CCAO-F vs CCDV-F vs CCAR-F vs CCAR-P13 min read ArticleFree CCAO-F Practice Questions 2026: 12 Worked Examples by Domain11 min read ArticleCCAR-F vs CCA-F: Registration and Pearson VUE Guide for Claude Architect Foundations (2026)19 min read

Exam Review

Key Facts: Claude Certified Developer - Foundations Exam

720/1000

Passing Score

Anthropic Exam Guide v1.0 (July 2026)

Exam Questions

Anthropic (MC/MR)

120 min

Exam Time

Anthropic

Exam Domains

Anthropic CCDV-F exam guide

$125

Exam Fee

Anthropic certification page

12 mo

Validity

Anthropic Exam Guide v1.0

CCDV-F requires a scaled score of 720 out of 1000 across 53 multiple-choice and multiple-response questions in 120 minutes. The exam costs $125, is delivered in English, is valid for 12 months, and is scheduled through Pearson VUE after Partner Academy registration.

Sample Claude Certified Developer - Foundations Practice Questions

Try these sample questions to review concepts for the Claude Certified Developer - Foundations exam. Each question includes a detailed explanation. Start the interactive quiz above for the full 164+ question experience with AI tutoring.

1A team is building a customer-support feature with Claude. They want the simplest reliable design and will only add agent complexity if needed. What should they start with?

A.A multi-agent swarm in which several agents debate every reply

B.A single well-prompted call with retrieval, adding steps if needed

C.An orchestrator-worker system with five specialized workers from day one

D.A fine-tuned support model trained before any prompt is written

Explanation: Anthropic's guidance on building effective agents is to start with the simplest design that works: many support use cases are handled by one optimized call with retrieval and good examples. Agentic loops are added only when the simpler pattern measurably falls short, which keeps the system cheaper, more transparent, and easier to evaluate.

2You need to classify each support ticket and send billing questions to one prompt, technical issues to another, and general questions to a third. Which workflow pattern best fits?

A.Parallelization, drafting all three category answers concurrently

B.An evaluator-optimizer loop that critiques each draft answer

C.Prompt chaining through billing, technical, and general steps

D.Routing, classifying the ticket to one specialized prompt

Explanation: Routing classifies an input once and dispatches it to a specialized downstream prompt or model. It fits when the categories are genuinely distinct, classification is reliable, and each path benefits from separate tuning.

3A coding assistant generates a draft, then a separate call critiques it against quality criteria, then refines it until the critique passes. Which pattern is this?

A.Orchestrator-worker, where a manager delegates subtasks to workers

B.Routing, dispatching the draft to a category-specific prompt

C.Evaluator-optimizer, looping generate and critique until it passes

D.Parallelization, drafting independent sections concurrently

Explanation: Evaluator-optimizer uses one call to generate and a separate call to judge against stated criteria, looping until the criteria are met. Keeping generation and evaluation in separate calls is what makes the critique useful.

4A research task can be split into independent section drafts that are later merged. Which pattern best fits, and what is a key risk?

A.Prompt chaining, whose main risk is an unnecessarily long sequential token bill

B.Parallelization by sectioning, risking inconsistent facts without a merge

C.Routing, whose main risk is misclassifying the whole document

D.Evaluator-optimizer, whose main risk is an unbounded critique loop

Explanation: Sectioning is a parallelization pattern: independent subsections are drafted concurrently and then combined. Without a deliberate merge or synthesis step, tone and facts drift between sections because no single call ever saw them all.

5When is an orchestrator-worker agent pattern most appropriate?

A.When every request follows the same fixed three-step pipeline

B.When subtasks emerge at runtime and a manager must delegate

C.When the job only needs a single classification label returned

D.When the whole job already fits in one call without any tools

Explanation: Orchestrator-worker fits open-ended work where the set of subtasks is not known before the request arrives. A central orchestrator plans, delegates to workers, and integrates their results.

6In the Claude Agent SDK, what best describes a typical agent loop?

A.One non-streaming Messages call that never returns any tool_use blocks

B.A cycle of reasoning, tool calls, tool results, and continuation

C.A set of offline batch jobs with no interactive tool feedback

D.A fixed stored procedure whose steps the model cannot influence

Explanation: An agent loop interleaves reasoning with tool use: the model emits tool_use blocks, your runtime executes them and returns tool_result blocks, and the model continues until it answers or a stop condition such as max turns or a timeout fires.

7A pipeline extracts entities with a fast model, enriches each entity with tools on a mid-tier model, then writes the final report on the strongest model. What is the main benefit?

A.It guarantees fewer total tokens than one call to the strongest model

B.It puts fast models on routine steps and the strongest on synthesis

C.It removes the need to handle tool errors at the enrichment step

D.It guarantees the final report will never need human review

Explanation: Per-stage model selection is a standard cost, latency, and quality trade-off: a small model handles mechanical extraction, a mid-tier model drives tool calls, and the strongest model is reserved for the hard synthesis where quality is most visible.

8A developer implements an agent that can call many tools. The agent sometimes loops indefinitely calling the same failing tool. What is the best first mitigation?

A.Remove every tool so that the model can only produce plain chat text

B.Add max turns, repeated-failure detection, and clear tool errors

C.Send every request to the largest available model instead

D.Turn off streaming so partial output cannot be rendered

Explanation: Production agents need explicit stop conditions plus useful tool feedback. Capping turns, detecting the same failing call repeating, and returning an error string that tells the model to try something else stops the loop without removing capability.

9When should you prefer a fixed prompt-chaining pipeline over a free-form agent loop?

A.When the steps are known and stable and need no runtime planning

B.When the application happens to be running on the smallest model

C.When the available tool schemas change unpredictably per request

D.When the model is expected to define new tools during the run

Explanation: A fixed chain is more predictable, easier to unit test, and usually cheaper than an open-ended loop that re-plans every turn. Choose it whenever the sequence of steps is known in advance and does not depend on intermediate results.

10An evaluator-optimizer loop is not improving quality after several iterations. What is the most likely design issue?

A.The responses are being streamed rather than returned in one piece

B.The evaluator lacks testable criteria or shares the generator's prompt

C.The critique step was not submitted through the asynchronous Batches API

D.The project CLAUDE.md file was not attached to the API request

Explanation: Evaluator-optimizer only works when the evaluator has explicit, checkable criteria and enough independence from the generator. Vague criteria, or a self-check that reuses the generator's own prompt, produce agreeable praise instead of actionable critique.

About the Claude Certified Developer - Foundations Exam

Claude Certified Developer - Foundations (CCDV-F) is Anthropic's developer certification for building and shipping Claude-powered applications. The exam covers Applications and Integration (33.1%), Model Selection and Optimization (16.8%), Agents and Workflows (14.7%), Prompt and Context Engineering (11.0%), Tools and MCPs (10.6%), Security and Safety (8.1%), Claude Code (3.1%), and Eval, Testing, and Debugging (2.6%).

Exam sponsor: Anthropic. The requirements and fees below concern the certification or admission exam, separate from our free practice resources.

Questions

53 questions

Time Limit

120 minutes

Passing Score

720/1000

Exam / Certification Fees

$125

Exam sponsor website

Fees, eligibility, and exam policies can change. Confirm them with the exam sponsor before applying or paying.

Official sources

Our practice resources: topics covered

We aim to reflect publicly available exam outlines and topic information in our study resources. Coverage, format, and difficulty may differ from the actual exam, and we cannot guarantee that every detail is accurate or current. Confirm exam requirements, fees, and policies with the official exam sponsor.

14.7%

Agents and Workflows

Build agent loops and multi-step workflows with the Claude Agent SDK, including tool-using agents, chaining, routing, parallelization, and when simpler single-call designs are enough

33.1%

Applications and Integration

Ship production Claude apps with the Messages API: tool use, streaming, extended thinking, prompt caching, Batch API, structured outputs, and service integration

3.1%

Claude Code

Configure Claude Code for developer workflows with CLAUDE.md, settings.json, permissions, and hooks

2.6%

Eval, Testing, and Debugging

Design evals, measure quality regressions, and debug tool-use and production failures in Claude applications

16.8%

Model Selection and Optimization

Choose Opus, Sonnet, or Haiku for capability, latency, and cost; optimize with caching, batching, streaming, and context strategies

11.0%

Prompt and Context Engineering

Design system prompts, examples, retrieval patterns, and context-window strategies for reliable developer workflows

8.1%

Security and Safety

Defend against prompt injection, enforce tool-permission boundaries, apply guardrails, and handle data safely

10.6%

Tools and MCPs

Define tool schemas, handle tool results, and integrate MCP servers (tools, resources, prompts) over stdio and HTTP

Preparing for the Claude Certified Developer - Foundations Exam

What You Need to Know

Passing score: 720/1000
Exam length: 53 questions
Time limit: 120 minutes
Exam / certification fees: $125 Official sources

Using Our Practice Resources

Work through all 164 available questions
Review every answer and explanation
Track weak areas and revisit them
Use our AI tutor for tough concepts

Claude Certified Developer - Foundations: Suggested Study Strategy

1Prioritize Applications and Integration (33.1%): Messages API, tool use, streaming, extended thinking, prompt caching, and the Batch API

2Know Opus vs Sonnet vs Haiku trade-offs for capability, latency, and cost — model selection is ~17% of the exam

3Practice agent workflow patterns: chaining, routing, parallelization, orchestrator-worker, and evaluator-optimizer — and when a single call is enough

4Master MCP primitives cold: tools, resources, and prompts, plus stdio vs HTTP transports

5Study prompt caching and context management as first-line cost and latency levers in production apps

6Treat security seriously: prompt injection, tool-permission boundaries, and guardrails appear throughout user-facing and agentic designs

7Know Claude Code basics: CLAUDE.md project memory, settings.json, and hooks — small weight but high signal

8Build a small eval set for any app you study so you can explain how you would detect quality regressions

9Complete 100+ practice questions across all eight domains and review every wrong answer until you can explain the API or design principle behind the correct choice

Frequently Asked Questions

What is the CCDV-F passing score?

CCDV-F requires a scaled score of 720 on a 100-1000 range. The exam has 53 multiple-choice and multiple-response questions in 120 minutes.

How much does the CCDV-F exam cost?

The current CCDV-F exam fee is $125 USD. Each retake is charged at the full exam fee.

What does the CCDV-F exam cover?

CCDV-F covers eight domains: Applications and Integration (33.1%), Model Selection and Optimization (16.8%), Agents and Workflows (14.7%), Prompt and Context Engineering (11.0%), Tools and MCPs (10.6%), Security and Safety (8.1%), Claude Code (3.1%), and Eval, Testing, and Debugging (2.6%).

What prior experience do I need for CCDV-F?

Anthropic lists no formal prerequisite. The exam guide targets hands-on developers with roughly 1–5 years of software engineering experience and at least six months building with the Claude API, typically in Python and/or TypeScript.

How long should I study for CCDV-F?

Most candidates need 3–10 weeks (roughly 40–80 hours) depending on prior Claude API experience. Prioritize Applications and Integration (about one-third of the exam), then model selection, agents/workflows, and tools/MCP. Anthropic Academy offers free aligned courses including Building with the Claude API, Claude Code 101, and Introduction to MCP.

How is CCDV-F different from CCAR-F?

CCDV-F is the developer exam focused on building and shipping Claude applications (Messages API, tools, caching, MCP integration). CCAR-F (Claude Certified Architect - Foundations) emphasizes production agentic architecture and multi-agent system design at a higher architecture altitude.

How is CCDV-F delivered and when can I retake it?

Register through the Anthropic Partner Academy, then schedule with Pearson VUE for online-proctored or test-center delivery. Failed attempts require 14-, 30-, then 90-day waits, with four attempts per rolling 12 months. The credential is valid for 12 months.