Claude Partner Network · The Gnar Company
Claude certification prep
Practice exams and blueprint reference for the Claude certifications, built from Anthropic's
published exam guides. 120 original practice items for CCAR-F; the full transcribed blueprint for 3 credentials.
Unofficial, and not affiliated with or endorsed by Anthropic. Every exam fact here is transcribed
from the official exam guide, which remains the authoritative reference — read it in full before
scheduling.
Choose a credential
Each description below is Anthropic's own, from the “Intended Audience” section of that
credential's exam guide.
CCAR-F
Claude Certified Architect – Foundations
120 practice items
The ideal candidate for this certification is a solution architect who designs and implements production applications with Claude.
- Exam
- 60 items · 120 min
- Fee
- $125
- Passing score
- 720 of 1000
- Blueprint
- 5 domains
· 30 task statements · 6 scenarios
Practise CCAR-F
Exam blueprint
CCDV-F
Claude Certified Developer – Foundations
reference only
The certification is intended for technical professionals who build, integrate, and ship production-grade AI solutions using large language models (LLMs), particularly Anthropic's Claude platform.
- Exam
- 53 items · 120 min
- Fee
- $125
- Passing score
- 720 of 1000
- Blueprint
- 8 domains
Exam blueprint
CCAR-P
Claude Certified Architect – Professional
reference only
The certification is intended for mid- to senior-level technical professionals who design, build, and deliver production-grade AI solutions using large language models, particularly Claude.
- Exam
- 63 items · 120 min
- Fee
- $175
- Passing score
- 720 of 1000
- Blueprint
- 7 domains
Exam blueprint
CCDV-F and CCAR-P have a fully transcribed blueprint and no practice items
yet — domains, weights, published objectives, fees, scoring and policies are all current, but
an original item bank has to be authored against them first.
CCAO-F is transcribed in this repo but earns no Partner Network Services Track credit, so it is not prepared here.
How practice runs work
- Drill
- 20 items · untimed · marked as you answer
Each answer is graded immediately, with the correct option, the reason it is correct, and
the named reasoning error behind the option you chose.
- Simulation
- 60 items · 120 min · graded at the end
The published exam form and time limit. Nothing is marked until you finish, then you get
percent-correct by domain and by task statement, and an item-by-item review.
Every incorrect option in the bank is tagged with one of five recurring
reasoning errors, so a run reports which one you make most often and can drill it. See the five categories — that
taxonomy is ours, not Anthropic's.
Scoring and sourcing
Practice runs report percent-correct on original items. The passing score is scaled from a
standard-setting study against the real item pool, so no practice percentage converts into it,
and nothing here predicts a pass.
Every practice item is written against the objectives published in the exam guides. None is
transcribed or reconstructed from a real exam, which would breach the certification agreement.
Exam guides are subject to change without notice. Re-verify each exam cycle; the guide is the
authority over anything shown here.