Claude Certified Developer Practice Exam

Q: How hard is the CCDV-F exam?

The exam guide targets an engineer with one to five years of experience and six months hands-on with Claude or comparable LLM systems. A third of the marks are applications and integration, so it rewards people who have shipped against the API; prompting-only experience is the common gap.

Q: Is this a full-length CCDV-F practice exam?

Yes: 53 original questions in 120 minutes, matching the real exam's count and clock, drawn to the published domain weights with a third of the questions on applications and integration.

Q: Is the CCDV-F practice exam free?

Yes. No sign-up. Results stay in your browser.

Q: How is the CCDV-F practice exam scored?

Percent correct mapped onto the 100 to 1,000 scale against the 720 pass mark, labeled as an estimate because the program's scaling is unpublished.

Q: Are these CCDV-F exam dumps?

No. They are original questions written from the exam guide and public API documentation. Using leaked exam content violates the candidate agreement.

Question 1

A response comes back with stop_reason set to max_tokens. What happened?

Accepted Answer

The output reached the max_tokens limit and was cut off before the model finished

Answer

The model reached the natural end of its turn, so the text is complete as returned

Answer

The model emitted a tool_use block and is waiting for a tool_result from your code

Answer

The input exceeded the context window and the request was rejected with a 400

Question 2

A response has stop_reason tool_use. What must your code do next?

Accepted Answer

Execute the requested tool, append a tool_result block with the matching id, and call the model again

Answer

Return the text blocks to the user, since the API already executed the tool on the server side before responding

Answer

Retry the same request with tool_choice set to none so the model answers in plain text

Answer

Wait for the model's next message, which will carry the tool_result once execution finishes

Question 3

Users should see the answer appear as it is generated. What do you use?

Accepted Answer

Set stream to true and render content_block_delta events as they arrive

Answer

Poll the messages endpoint with the message id until a status field reports completion

Answer

Submit the request through the Message Batches API and read results as they land

Answer

Lower max_tokens so each response returns quickly enough to feel incremental

Question 4

You have 20,000 documents to classify overnight, with no latency requirement, and cost matters. Best choice?

Accepted Answer

Submit them through the Message Batches API and collect results when the batch completes

Answer

Send 20,000 synchronous requests through a worker pool sized to your rate limit

Answer

Stream each request so results can be written to disk as the tokens arrive

Answer

Concatenate all the documents into one request with a large max_tokens and split the output back apart by document

Question 5

Every request includes the same 30,000-token reference manual followed by a short question. How do you cut cost?

Accepted Answer

Put a cache_control breakpoint on the manual block so the prefix is reused across requests

Answer

Send the question first and the manual second so the model reads the task before the context

Answer

Upload the manual through the Files API so its tokens are billed once instead of per request

Answer

Lower temperature and max_tokens so each request generates fewer billable output tokens

Question 6

Your integration starts receiving HTTP 429 responses. What is the right client behavior?

Accepted Answer

Back off exponentially with jitter and honor the retry-after header when it is present

Answer

Retry immediately in a tight loop, since the limit clears as soon as the window rolls over

Answer

Rotate to a second API key from the same organization so requests keep flowing

Answer

Treat 429 as a transient server fault and fail the request back to the user

Question 7

A pipeline extracts three fields from short emails, millions of times a month. Which model tier makes sense?

Accepted Answer

The smallest tier that passes your eval, since cost and latency dominate at this volume

Answer

The most capable tier, since extraction errors at this scale compound quickly

Answer

The tier with the largest context window, so that long quoted email threads are not truncated before extraction

Answer

A mid tier with extended thinking enabled to raise accuracy on ambiguous fields

Question 8

You need the same input to produce the same extraction every time. Which setting helps most?

Accepted Answer

Set temperature near 0 so sampling favors the most likely tokens

Answer

Set temperature near 1 so the model commits firmly to a single answer

Answer

Enable prompt caching so the cached prefix yields the same output

Answer

Add a longer system prompt that describes each field in detail

Question 9

Which loop correctly implements a tool-using agent?

Accepted Answer

Call the model; on tool_use run the tool, append the result, call again; stop on end_turn or a turn cap

Answer

Run every declared tool up front with default arguments, then pass all of the results to the model in a single call

Answer

Call the model once with tool_choice any and return the first tool_use arguments

Answer

Call the model repeatedly until its text no longer mentions any tool by name

Question 10

What does the Claude Agent SDK handle for you that a hand-rolled loop does not?

Accepted Answer

The agent loop, tool execution, permission checks, hooks, and subagents

Answer

Automatic retries, rate-limit backoff, and typed request and response objects

Answer

Server-side tool execution so your code does not have to handle tool_use blocks

Answer

Fine-tuning the model on your repository before the first run

Question 11

A multi-turn assistant's history grows until requests fail. What is the sane fix?

Accepted Answer

Summarize older turns and keep the most recent turns verbatim

Answer

Send only the latest user message and rely on the system prompt to carry state

Answer

Add a cache_control breakpoint so older turns stop counting toward the window

Answer

Raise max_tokens so the request budget covers the longer history

Question 12

A tool parameter may only be one of three values. How do you express that?

Accepted Answer

Use an enum in the parameter's JSON schema and list the parameter under required

Answer

Describe the three allowed values in the tool description text and validate the string in your handler

Answer

Split the capability into three separate tools, one per allowed value

Answer

Set tool_choice to the tool's name so the model cannot pass other values

	Real exam	This practice exam
Questions	53	53
Time	120 minutes	120 minutes
Domain weights	Per the published guide	Same weights, applied when the set is drawn
Scoring	100 to 1,000, 720 to pass	Same scale, linear estimate
Format	Proctored, closed-book, no AI	Unproctored; treat it as closed-book to get a real reading
Question bank	Anthropic's, confidential	60 original questions, fresh draw each attempt

The free Claude Certified Developer (CCDV-F) practice exam: 53 questions, 120 minutes.

Take the CCDV-F practice exam

Full practice exam

Quick check

How closely it matches the real CCDV-F

Domain weights used for the draw

How to use it

Take a quick check first

Study the weak domains, then sit the full timed exam

Review every miss, then book when you clear the bar

CCDV-F mock exam, practice test, or practice questions?

What to check in any practice exam

CCDV-F practice exam FAQ

Claude Certified Architect – Foundations

Claude Certified Associate – Foundations

Claude Certified Architect – Professional

Who wrote this

Matei Olaru

Michael Greenberg