About this book and the exam
CCAR-P is Anthropic's senior credential for architects who design, integrate, govern, and run Claude systems at enterprise scale. This book follows the official v1.0 exam guide (effective July
- 2026) domain by domain, so every chapter maps to a weighted slice of the exam.
Who it is for
The exam guide recommends 3+ years in systems architecture or platform engineering, 6+ months of hands-on production work with Claude or a comparable LLM, and experience shipping end-to-end systems. It states the credential is not intended for entry-level developers. There is no mandatory prerequisite, but most candidates take Architect Foundations (CCAR-F) first, and this book assumes you know that material.
- Exam facts
- Item Detail Status
- Code
- and ver‐
- sion
CCAR-P, exam guide v1.0, effective July 2026 Verified in official guide Items 63 multiple-choice and multiple-response; each item says how many answers to pick
Verified in official guide
Time 120 minutes (plan about 135 minutes of seat time with check-in) Verified; seat time is a third-party estimate
Scoring Scaled 100 to 1,000, pass at 720, criterion-referenced Verified in official guide Score re‐
port
Pass/fail, scaled score, percent correct by domain Verified in official guide Fee $175 USD; partner-tier discounts reported Fee verified; discounts from third-party reporting
Delivery Proctored, online (OnVUE) or Pearson VUE test center, closed book, English and Japanese; retakes after 14, 30, then 90 days,
- max 4 attempts per 12 months; Credly badge
- Proctoring verified in exam guide; the
- rest per Anthropic's partner FAQ and
- Pearson VUE page
Validity 12 months; free non-proctored renewal reported 12 months verified; renewal method unconfirmed
Eligibility Registration reportedly needs a work email at a Claude Partner Network organization
Figure 1. The seven CCAR-P domains by exam weight
Domain Weight About items Chapter
- 3. Integration 19% 12 Chapter 4
- 1. Solution Design & Architecture 17% 11 Chapter 2
- 4. Evaluation, Testing & Optimization 16% 10 Chapter 5
- 5. Governance, Safety & Risk Management 14% 9 Chapter 6
- 6. Stakeholder Communication & Lifecycle Management 14% 9 Chapter 7
- 2. Claude Models, Prompting & Context Engineering 13% 8 Chapter 3
- 7. Developer Productivity & Operational Enablement 7% 4 Chapter 8 Item counts are estimates from the weights. Domains 5, 6, and 7 (35% together) have no counterpart on the Foundations exam, so budget extra study time there.
How to use this book
Read Chapter 1 for the mindset the exam rewards.
Work through Chapters 2 to 8. Each opens with what the exam tests and the official task statements, teaches the concepts, then gives Common traps and Check yourself questions. Use Chapter 9 as your reference architecture and Chapter 10 for exam-day tactics. Do the six capstone cases, sit the 40-question Practice exam 1 under a 75-minute timer, then sit the full-length 63-item Practice exam 2 under real exam conditions (120 minutes). Work through the labs in Chapter 12 and the rapid review in Chapter 13, and keep Practice exam 3 as your final mock in the last week.
Score 80% or better twice before booking.
- 3. Integration19%
- 1. Solution Design & Architecture17%
- 4. Evaluation, Testing & Optimization16%
- 5. Governance, Safety & Risk Management14%
- 6. Stakeholder Communication & Lifecycle14%
- 2. Claude Models, Prompting & Context13%
- 7. Developer Productivity & Enablement7%
Domains 5, 6 and 7 (35% together) have no counterpart on the Foundations exams.
Product details (model names, prices, feature names) change quickly. Where this book quotes one, treat it as accurate at the date above and confirm it in Anthropic's documentation.