Which Certification?
A decision framework for choosing between CCAO-F, CCDV-F, CCAR-F and CCAR-P, with side-by-side blueprints.
Two questions settle it: do you write code most days? and do you design systems or build them?
Decision tree
Do you write integration code (API, SDK, MCP, agents) as part of your job?├─ No → Do you advise, sell, manage, analyse or operate with Claude?│ └─ Yes → CCAO-F (Associate – Foundations)└─ Yes → Do you mainly BUILD what others design, or DESIGN what others build? ├─ Build → CCDV-F (Developer – Foundations) └─ Design → Do you own discovery → design → handoff → monitoring → iteration, including stakeholder communication, RAG and governance? ├─ Not yet → CCAR-F (Architect – Foundations) └─ Yes → CCAR-P (Architect – Professional) (Anthropic recommends CCAR-F first unless you have deep, current production experience)Side-by-side blueprints
CCAO-F – Associate – Foundations (60 items, $99)
| # | Domain | Weight |
|---|---|---|
| 1 | Prompting and Task Execution | 14% |
| 2 | Output Evaluation and Validation | 21% |
| 3 | Product and Model Selection | 12% |
| 4 | Workflow Integration and Solution Design | 16% |
| 5 | Configuration and Knowledge Management | 12% |
| 6 | Governance, Risk, and Responsible Use | 15% |
| 7 | Troubleshooting and Optimization | 10% |
Evaluation, workflow design and governance together exceed half the exam. It tests judgment about using AI responsibly more than prompting technique.
CCDV-F – Developer – Foundations (53 items, $125)
| # | Domain | Weight |
|---|---|---|
| 1 | Applications and Integration | 33.1% |
| 2 | Model Selection and Optimization | 16.8% |
| 3 | Agents and Workflows | 14.7% |
| 4 | Prompt and Context Engineering | 11.0% |
| 5 | Tools and MCPs | 10.6% |
| 6 | Security and Safety | 8.1% |
| 7 | Claude Code | 3.1% |
| 8 | Eval, Testing, and Debugging | 2.6% |
The top three domains cover almost two-thirds of the items. Claude Code and Eval together are under 6%.
CCAR-F – Architect – Foundations (60 items, $125)
| # | Domain | Weight |
|---|---|---|
| 1 | Agentic Architecture and Orchestration | 27% |
| 2 | Claude Code Configuration and Workflows | 20% |
| 3 | Prompt Engineering and Structured Output | 20% |
| 4 | Tool Design and MCP Integration | 18% |
| 5 | Context Management and Reliability | 15% |
Questions are drawn from 4 of 6 published scenarios per sitting. Candidates report Agentic Architecture as the hardest domain.
CCAR-P – Architect – Professional (63 items, $175)
| # | Domain | Weight |
|---|---|---|
| 1 | Solution Design and Architecture | 17% |
| 2 | Claude Models, Prompting and Context Engineering | 13% |
| 3 | Integration (incl. RAG) | 19% |
| 4 | Evaluation, Testing and Optimization | 16% |
| 5 | Governance, Safety and Risk Management | 14% |
| 6 | Stakeholder Communication and Lifecycle Management | 14% |
| 7 | Developer Productivity and Operational Enablement | 7% |
All five CCAR-F domains reappear at higher altitude inside domains 1–5. Domains 3 (RAG), 4 (evals/A-B testing), 5 (regulated compliance) and 6 (stakeholders/lifecycle) are the deltas to study on top.
Intended and not-intended audiences
| Credential | Intended for | Explicitly not intended for |
|---|---|---|
| CCAO-F | Operations, marketing, PM, education, communications, general knowledge work; limited-to-moderate technical expertise | Developers building against APIs or designing agentic systems |
| CCDV-F | AI/ML engineers, technical leads, senior software engineers; 1–5 yrs SWE, 6+ months with Claude, Python/TypeScript | Non-technical users; roles limited to prompt writing |
| CCAR-F | Solution architects with 6+ months on Claude APIs, Agent SDK, Claude Code, MCP | Those without system-design responsibility |
| CCAR-P | Mid-to-senior solution architects, AI/ML engineers, technical leads; 3+ yrs architecture, 6+ months Claude in production | Entry-level developers, casual users, isolated-task roles |
Common pairings
- Developer + Architect Foundations ($250): the most common technical pairing; heavy overlap in agents, tools/MCP, Claude Code and context management, so study time is shared.
- Associate first, then a technical track: good for consultants moving toward solution work.
- Architect Foundations → Professional: the intended ladder. Professional assumes Foundations-level fluency and adds lifecycle, RAG and governance depth.
There is no ladder requirement
You may sit any exam in any order. Failing one exam does not block registration for a different one. But the retake waits (14/30/90 days) and full fees per attempt make benchmarking with a practice exam before booking the cheapest decision you will make.
Last updated Sep 18, 2026