Claude Certification Blog
CCAR-P vs CCAR-F: what actually changes at the Professional level
CCAR-P vs CCAR-F is not a matter of a few harder questions. Foundations has five domains, 60 items and a $125 fee; Professional has seven domains, 63 items, a $175 fee, and broadens the role to end-to-end architecture, governance, stakeholders and operations.
CCAR-P vs CCAR-F is a choice between two Architect credentials, not a prerequisite chain: the current CCAR-P guide lists no mandatory prerequisites and does not require CCAR-F. The old domain-direction story is now backwards: CCAR-F has five domains and 60 items at $125; CCAR-P has seven domains and 63 items at $175. Both allow 120 minutes and use a 720 scaled passing score, but those shared logistics do not establish relative difficulty.
CCAR-P vs CCAR-F: the nine-row decision table
Every cell below comes from the two version 1.0 exam guides or from the programme dates published on the CCAR-F and CCAR-P certification pages. Confirm fees and dates against the official guide before booking; the partner-tier discount is applied at checkout.
| Decision row | CCAR-F · Foundations | CCAR-P · Professional |
|---|---|---|
| Best for | Solution architects who build Claude systems hands-on: Agent SDK loops, Claude Code, MCP tools, structured output | Mid- to senior-level practitioners who own end-to-end solution design, integration, evaluation, governance and stakeholders |
| Level | Foundations | Professional |
| Focus | Five implementation domains; each sitting draws four of six published scenarios | Seven lifecycle domains, from discovery through design to monitoring and iteration |
| Length | 120 minutes | 120 minutes |
| Questions | 60 items, 720 of 1,000 to pass | 63 items, 720 of 1,000 to pass |
| Price | $125 per attempt ($99 until the 30 June 2026 move to Pearson VUE) | $175 per attempt |
| Validity | 12 months; free non-proctored renewal if on time | 12 months; free non-proctored renewal if on time |
| Delivery | Pearson VUE, online proctored or test centre, since 30 June 2026; launched 12 March 2026 on Skilljar | Pearson VUE, online proctored or test centre; announced 23 July 2026 |
| Prerequisites | None mandatory; 6+ months building with Claude recommended | None mandatory and CCAR-F is not required; 3+ years of architecture and 6+ months with Claude recommended |
The current Architect – Foundations page focuses on agentic architecture, MCP/tool design, Claude Code workflows, prompt/structured-output work and context reliability, framed by four of six published production scenarios per sitting. The Architect – Professional page broadens that lens to end-to-end solution design, enterprise integration, evaluation, governance and risk, stakeholder and lifecycle work, and developer enablement.
Which first: CCAR-F or CCAR-P?
There is no required order, so the question is about your job, not the programme. If your week is spent building the systems (agent loops, Claude Code configuration, MCP tools, extraction schemas), the Foundations paper examines exactly that work and is the natural first sitting. If you already own end-to-end delivery, run discovery with stakeholders, and are accountable for evaluation and governance, the Professional guide describes your role directly, and Anthropic does not make you pass Foundations to prove it.
Two facts should weigh on the order. Fees are per attempt and retakes cost the full fee, so sitting the paper whose task statements you can already evidence is cheaper than sitting both in hope. And CCAR-P is the newer credential, announced on 23 July 2026, while CCAR-F launched on 12 March 2026 and moved to Pearson VUE on 30 June 2026. If the choice is really between all four exams rather than these two, the guide to choosing a Claude certification owns that decision.
Which is harder? "Professional" is a role level, not a difficulty formula
Anthropic labels CCAR-F as Foundations and CCAR-P as Professional. The Professional guide targets mid- to senior-level technical practitioners and recommends three or more years in systems architecture or platform engineering, six or more months with Claude or comparable production LLM systems, and end-to-end delivery experience. It also says those are recommendations, not prerequisites.
That profile supports a role comparison, not a promise that one test will feel harder or easier. The two papers also make different demands: CCAR-F holds you inside four production scenarios for the whole sitting, so the cost of misreading one context is spread across many items; CCAR-P spreads 63 items across seven domains, so breadth matters more than depth in any one. A person's job history, recent practice and blueprint gaps decide which of those feels harder.
Where the current blueprint weight sits
CCAR-F's largest domain is Agentic Architecture & Orchestration at 27%. CCAR-P is more evenly distributed: Integration is largest at 19%, followed by Solution Design & Architecture at 17% and Evaluation, Testing & Optimization at 16%. Its two largest domains total 36%, not 60%.
The two current blueprints, side by side
The rows below preserve each guide's published order; they are not one-to-one equivalences. The table makes the change in structure visible without forcing unlike domains into a false mapping.
| CCAR-F · Foundations | Weight | CCAR-P · Professional | Weight |
|---|---|---|---|
| Agentic Architecture & Orchestration | 27% | Solution Design & Architecture | 17% |
| Tool Design & MCP Integration | 18% | Claude Models, Prompting & Context Engineering | 13% |
| Claude Code Configuration & Workflows | 20% | Integration | 19% |
| Prompt Engineering & Structured Output | 20% | Evaluation, Testing & Optimization | 16% |
| Context Management & Reliability | 15% | Governance, Safety & Risk Management | 14% |
| None | None | Stakeholder Communication & Lifecycle Management | 14% |
| None | None | Developer Productivity & Operational Enablement | 7% |
What can transfer from Foundations?
Transfer is real, but it is evidence-based rather than automatic. The following is a Cred Farmer mapping of related tasks in the two current guides, not an official prerequisite ladder or a claim that passing one domain proves mastery of another.
| CCAR-F evidence | How CCAR-P applies it |
|---|---|
| Agentic architecture and orchestration | Compare end-to-end patterns, multi-agent strategies and business constraints. |
| Tool design and MCP integration | Evaluate protocols, permissions, capability bloat, observability and integration risk. |
| Prompts, structured output and context reliability | Design model/prompt/context strategy, evaluation criteria and optimization loops. |
| Claude Code configuration and workflows | Enable developer teams, improve workflows and support operational debugging. |
CCAR-P also gives Governance, Safety & Risk Management and Stakeholder Communication & Lifecycle Management their own 14% domains. CCAR-F experience can support those decisions, but its blueprint gives neither a domain of its own: see what the stakeholder domain examines, and a fortnight that front-loads it rather than the heaviest domain.
How to prepare for CCAR-P after CCAR-F
Do not assume a shorter plan because you hold CCAR-F: Anthropic publishes no conversion from one credential to a preparation time for the other. Size the work from the Professional task statements and your own evidence.
Download the current Professional guide
Do not plan from an old five-domain snapshot. Mark every current task statement as proven, partial or unproven.
Map evidence, not domain names
For each prior skill, write how it supports a Professional task and where it stops. Similar vocabulary is not automatic mastery.
Practise full-lifecycle decisions
Write short architecture briefs that include requirements, trade-offs, evaluation, security, governance, stakeholders and operations.
Use fresh items for the final signal
Repair gaps by domain before using unseen full-length mocks; keep repeated or untimed attempts as narrower learning evidence.
For a cold read on either paper, the free timed CCAR-P mock runs the 63-item form with a reason for every option, and the free timed CCAR-F set is a 20-item paper across the five official domains, not a replica of the scenario-based exam. The Claude certification study guide provides the practice loop, while the mock-test guide explains why fresh forms should be saved for readiness checks. If the question is whether the second badge has enough role value to justify its $175 fee, use the worth-it decision framework.
CCAR-F is optional; the Professional profile is not decorative
CCAR-P has no mandatory prerequisite, but its guide still targets experienced practitioners who design, implement and govern end-to-end production AI systems. Use that profile to judge fit even when registration does not enforce it.
Checked against the official CCAR-F and CCAR-P exam guides (version 1.0, July 2026) and the Anthropic Partner Academy certification pages on 17 September 2026.
Key takeaways
- CCAR-F is not a CCAR-P prerequisite. The Professional guide lists no mandatory prerequisites or courses.
- The current structure is five domains versus seven. Any seven-to-five comparison is using an obsolete blueprint.
- $125 versus $175, per attempt. CCAR-F rose from $99 when it moved to Pearson VUE on 30 June 2026; retakes are charged in full on both.
- Same time and cut score do not prove equal difficulty. Compare the role profiles and task statements instead.
- Prior knowledge can transfer without guaranteeing readiness. Map evidence task by task and repair the uncovered Professional domains.
Review the current Professional blueprint
Check all seven domains and their task statements, then build a gap plan from evidence rather than an old domain map.
See the CCAR-P blueprintQuestions
Frequently asked
The follow-up questions people search next.
Is CCAR-P harder than CCAR-F?
Anthropic does not publish a difficulty conversion between the two exams. CCAR-P is labelled Professional and targets mid- to senior-level technical practitioners with end-to-end architecture experience; CCAR-F is labelled Foundations and targets solution architects implementing Claude systems. The same 120-minute limit and 720 scaled cut score do not prove equal difficulty.
Do you need CCAR-F before CCAR-P?
No. The current CCAR-P guide says there are no mandatory prerequisites or courses and does not list CCAR-F as a requirement. Its recommended experience (systems architecture, platform engineering and production LLM delivery) is still important role-fit guidance.
How much do CCAR-F and CCAR-P cost?
CCAR-F is $125 per attempt, up from $99 when delivery moved to Pearson VUE on 30 June 2026; CCAR-P is $175 per attempt. Partner-tier discounts apply at checkout, retakes are charged in full each time, and both credentials are valid for 12 months with a free on-time renewal.
What carries over from CCAR-F to CCAR-P?
Architecture, agent/tool integration, prompting, context, reliability and Claude Code experience can support Professional tasks. Transfer is not one-to-one: CCAR-P asks candidates to connect those skills to end-to-end design, evaluation, governance, stakeholder and operational decisions.
How long should CCAR-P preparation take after CCAR-F?
Anthropic publishes no preparation-time estimate or shortcut for CCAR-F holders. Use the current Professional task statements and a short mixed diagnostic to identify evidence gaps; then plan by gaps rather than assuming the second exam will take less time.
Keep reading
Related posts
Not affiliated with, or endorsed by, Anthropic or Pearson VUE. Details are summarised from publicly published program information and can change. Always confirm against the official exam guide before booking.