AI Certificates

Claude Certified Architect — Professional

Passed · 965/1000

Claude Certified Architect — Professional practice exams

6 sets, 332 questions, every option explained. Architecture, integration, evaluation, and governance for Claude systems. Four full-length 63-question forms plus two focused drills, every answer option explained.

Form 1 is free, no account.

Get full access $39

All 828 questions across the four certifications.

Verified buyer · Kiril Stoichkov
“The material was clear, well-organized, and gave me real confidence going into the exam. The platform itself was also easy and pleasant to use — simple to navigate, and it made studying feel less like a chore.”
Items per form
63
Fee
$175 USD
Pass mark
720 / 1000
Valid for
12 months

Booking the real exam needs a partner company email. No partner? See how sponsorship works

Start here

Form 1 is free — all 63 questions

No account, no email. Every option explained as you go.

Quick round10 questions, ~12 minutes — one from every domain, timed.Start →
Want the other 5 sets? $39 unlocks all 4 Claude certifications — 828 questions.
See full access

What the exam weights

D3 Integration · 19% of the exam · 70 questions in the bank · 12 on the free form

The 6 sets

Two ways to unlock it
Everything, one price$39
Full access · all four Claude certifications
CCAR-PArchitect — Professional269
CCAR-FArchitect — Foundations60
CCDV-FDeveloper — Foundations239
CCAO-FAssociate — Foundations260
828 questions total, on this site and in Claude or ChatGPT.
Get full access
Just this exam$19
CCAR-P only
All 6 CCAR-P sets, on this site and in Claude or ChatGPT. Nothing for the other three certifications.
Get CCAR-P only
Not ready to pay? Take the free set — 63 questions, no account.

The full study guide

Free · no signup · 7 partsEvery domain on the blueprint, written outThe official course material sits behind a partner login. This covers the same ground, written to the published blueprint by someone who sat the exam and scored 965 out of 1000, with all 63 free practice questions worked through in the sections they belong to.Start with the overview →

The question most people get wrong

D2 Claude Models, Prompting & Context EngineeringFrom the free form

A legal-tech firm analyzes merger agreements for cross-document inconsistencies: about 200 requests per day, each carrying roughly 150K tokens of contract text, with attorneys reviewing every output. Errors that reach attorneys are expensive to catch, users tolerate multi-minute turnaround, and daily model spend at current volume is under $900 even on the most capable model. The team is debating whether to cut costs with a smaller model. What should they do?

  1. A

    Stay on the most capable model, since at 200 requests/day the absolute cost is small relative to the cost of missed inconsistenciesCorrect

  2. B

    Switch to the smallest model and compensate with a far more detailed system prompt describing the legal reasoning steps to follow

  3. C

    Split each 150K-token contract into 10K-token chunks and run them in parallel through a mid-tier model to cut per-request cost and turnaround time

  4. D

    Route straightforward merger agreements to a small model and complex ones to the most capable model, using contract length as the routing heuristic the team maintainsMost picked

60.5% of 304 people got this wrong in the last 30 days (as of 2026-09-15). The most popular answer was D, chosen by 51 of the 96 who answered, and every one of them was wrong.

Why D fails

Tiered routing is sound engineering, but document length is a poor proxy for analytical difficulty and adds routing complexity to save a few hundred dollars a day.

The principle

Cost optimization only matters when cost is actually the binding constraint. When volume is low, error costs are high, and latency is slack, the correct trade is the most capable model, and premature downgrading or chunking optimizes an axis nobody is constrained on.

See how you do on the rest →

How to prepare for CCAR-P

  1. 1Sit Form 1, free and without an account. Too long for today? The quick round is ten questions, one per domain.
  2. 2Read the explanation under every option, the wrong ones most of all. The hard questions on this exam turn on why the tempting answer is wrong.
  3. 3Check the result page for your weakest domain and objective, then take the readiness check for a second reading against everyone else who sat it.
  4. 4Study the guide for that domain. The 11 guides below are grouped by what the exam weights. Then check your reasoning against the CCAR-P cheat sheet, which sets each domain's principles beside the traps that cost marks.

A worked question, every option explained

D1 Solution Design & ArchitectureFrom the free CCAR-P form

Nordbank Wealth generates quarterly suitability letters for 30,000 clients: fetch portfolio data, apply a fixed set of disclosure rules, and render the letter in a regulator-approved format. The steps are identical for every client, and auditors must be able to trace exactly what happened at each stage. Which architectural pattern best fits?

  1. A

    An autonomous agent with tool access that decides at runtime which data to fetch and how to compose each letter.

    Agent autonomy adds run-to-run variability and cost precisely where the process demands determinism and auditability, buying flexibility the task never uses.

  2. B

    A single comprehensive prompt containing the portfolio data, the disclosure rules, and formatting instructions in one Claude call.

    A single monolithic call works at small scale but gives auditors no intermediate artifacts and makes one failure mode indistinguishable from another.

  3. C

    A prompt-chained workflow with fixed, ordered steps and validation checkpoints between each stage of the letter pipeline.Correct

    Fixed steps with inter-stage checkpoints match a fully predictable process, and each stage's inputs and outputs create the audit trail regulators require.

  4. D

    A multi-agent system where specialist agents for data, compliance, and drafting negotiate the final letter.

    Multiple negotiating agents maximize coordination overhead and nondeterminism for a task with zero runtime ambiguity to resolve.

The principle

Pattern selection follows task predictability: when the steps are known in advance and identical every run, a structured workflow beats agentic autonomy on cost, latency, reliability, and auditability. Reserve agents for problems whose path cannot be predetermined.

Written by Matthew Hartman, who holds the Claude Certified Architect — Professional credential, passed with 965 out of 1000 in July 2026. Independent practice material, not affiliated with or endorsed by Anthropic.

11 free guides

Practice inside your assistant

Claude and ChatGPT can quiz you on CCAR-P — without being able to give you the answer

All 63 free CCAR-P questions connect to the assistant you already use. It gets the question and not the answer key, so it cannot hand you the answer or be talked into agreeing with you. You commit first; then it explains why each option wins or loses.

How it works

Also here