Anthropic has expanded its Claude certification programme into a set of four role-based credentials. For anyone building a practice around Claude, or working with a firm that does, these are worth understanding properly rather than at headline level. This post breaks down what each certification is, who it is for, how the exams are structured, and how to prepare, drawn directly from the official exam guides.

The short version

There are now four certifications, spanning two levels:

Certification	Code	Level	Fee (USD)	Items	Time
Claude Certified Associate – Foundations	CCAO-F	Foundations	$99	60	120 min
Claude Certified Developer – Foundations	CCDV-F	Foundations	$125	53	120 min
Claude Certified Architect – Foundations	CCAR-F	Foundations	$125	60	120 min
Claude Certified Architect – Professional	CCAR-P	Professional	$175	63	120 min

All four are proctored exams delivered through Pearson VUE, online or at a test centre. All use a scaled score of 100 to 1,000 with a cut score of 720 to pass. All credentials are valid for 12 months.

The exams sit behind the Claude Partner Network: exam access runs through a partner organisation, though the preparation courses on Anthropic’s Academy are free and open to anyone with an email address.

How the programme is structured

The four credentials map to three roles, at two levels of depth.

The Associate credential is the only non-technical one. It is aimed at people who use Claude as a productivity tool rather than people who build with it. The Developer and Architect – Foundations credentials both assume real engineering experience. The Architect – Professional credential sits above Foundations, aimed at senior practitioners leading enterprise delivery.

A sensible way to read the stack: Associate is the entry point for delivery and knowledge-work roles; Developer and Architect – Foundations are parallel technical credentials for people who already build with Claude; Architect – Professional is the step up for those designing and governing production systems end to end.

There are no mandatory prerequisites for any of the four. Each credential is awarded on exam performance alone. The recommended experience below is guidance, not a gate.

---

Claude Certified Associate – Foundations (CCAO-F)

Who it is for. Professionals who use Claude as a productivity tool in roles such as operations, marketing, project management, education, and communications. It covers using built-in platform features to streamline work, spotting where Claude can improve a process, balancing quality against cost and effort, and knowing when to escalate more technical work to a Developer or Architect.

This is explicitly not aimed at software developers or people designing agentic systems. It sits between the casual prompt user and the technical practitioner.

Recommended experience. Regular hands-on use of Claude in a professional setting, a grasp of structured problem-solving and workflow design, and a practical understanding of AI limitations such as hallucination, context limits, and data sensitivity. No development or API experience is needed.

Exam. 60 items, 120 minutes, $99.

What it tests (domain weights):

Domain	Weight
Output Evaluation and Validation	21%
Workflow Integration and Solution Design	16%
Governance, Risk, and Responsible Use	15%
Prompting and Task Execution	14%
Product and Model Selection	12%
Configuration and Knowledge Management	12%
Troubleshooting and Optimisation	10%

The largest single area is evaluating and validating Claude’s output: checking for accuracy, spotting hallucinations and bias, fact-checking, and deciding when human review is required. Governance and responsible use is weighted heavily too, covering data sensitivity, privacy, and appropriate versus inappropriate use cases.

---

Claude Certified Developer – Foundations (CCDV-F)

Who it is for. Engineers who build, integrate, and ship production applications, agents, and workflows on Claude. It covers building agents with the Claude Agent SDK, integrating Claude through the API and client SDKs, operating Claude Code, prompt and context engineering, evaluation, model selection, security, and building custom tools and MCP servers.

Recommended experience. One to five years in software engineering, at least six months hands-on with Claude or comparable LLM systems, proficiency in Python or TypeScript, fluency with REST APIs and CLI tools, and a working understanding of LLM fundamentals, agents, context management, and MCP.

Exam. 53 items, 120 minutes, $125.

What it tests (domain weights):

Domain	Weight
Applications and Integration	33.1%
Model Selection and Optimisation	16.8%
Agents and Workflows	14.7%
Prompt and Context Engineering	11.0%
Tools and MCPs	10.6%
Security and Safety	8.1%
Claude Code	3.1%
Eval, Testing, and Debugging	2.6%

The exam is weighted firmly toward applications and integration: API mechanics, software engineering foundations, application design across interfaces, and configuration management make up a third of it. Model selection and cost or token optimisation is the next largest area.

---

Claude Certified Architect – Foundations (CCAR-F)

Who it is for. Solution architects who design and implement production applications with Claude. It assumes hands-on experience building agentic applications with the Agent SDK, configuring Claude Code for team workflows, designing MCP tool and resource interfaces, engineering prompts for reliable structured output, managing context across long interactions, and integrating Claude into CI/CD pipelines.

Recommended experience. Six or more months building with the Claude API, Agent SDK, Claude Code, and MCP, with a working understanding of both the capabilities and the limits of LLMs in production.

Exam. 60 items, 120 minutes, $125. This is the one scenario-based exam of the four: it presents four scenarios drawn at random from a bank of six, each framing a set of questions around a realistic production context (a customer support agent, a multi-agent research system, Claude Code in CI/CD, structured data extraction, and so on).

What it tests (domain weights):

Domain	Weight
Agentic Architecture and Orchestration	27%
Claude Code Configuration and Workflows	20%
Prompt Engineering and Structured Output	20%
Tool Design and MCP Integration	18%
Context Management and Reliability	15%

Agentic architecture is the heaviest domain: agentic loops, coordinator-subagent patterns, task decomposition, and enforcement and handoff patterns. The exam rewards practical judgement about architecture and tradeoffs, not definitions.

---

Claude Certified Architect – Professional (CCAR-P)

Who it is for. Mid to senior technical professionals who design, build, and deliver production-grade Claude solutions and lead the decisions around them. Beyond the Foundations material, it adds retrieval-augmented generation, observability at scale, compliance with regulations such as GDPR and HIPAA, and the stakeholder and lifecycle side of delivery: discovery, communicating architectural tradeoffs, and managing expectations and SLAs.

Recommended experience. Three or more years in systems architecture or platform engineering, six or more months with Claude in production, and experience delivering end-to-end systems from discovery through operation.

Exam. 63 items, 120 minutes, $175.

What it tests (domain weights):

Domain	Weight
Integration	19%
Solution Design and Architecture	17%
Evaluation, Testing, and Optimisation	16%
Governance, Safety, and Risk Management	14%
Stakeholder Communication and Lifecycle Management	14%
Claude Models, Prompting, and Context Engineering	13%
Developer Productivity and Operational Enablement	7%

The weighting is noticeably flatter than Foundations. It spreads across integration, solution design, evaluation, governance, and stakeholder management, which reflects the breadth expected of someone owning a solution end to end rather than a single build.

---

How to prepare

Every exam guide says the same thing about preparation: there is no single required course, and no resource guarantees a pass. The route that works is self-study against the published blueprint plus genuine hands-on practice.

Three steps apply across all four:

1. Work the blueprint. The domain tables above are the map. Self-assess honestly against each domain and spend your time where you are weakest, weighted by how much of the exam each domain represents.
2. Build something real. For the technical credentials, build and operate at least one Claude application that exercises the heavy domains: for Developer and Architect, that means an agent with the SDK, at least one integrated tool or MCP server, structured output with JSON schemas, and basic evaluation and security. For Associate, configure a Claude Project with instructions and knowledge sources and practise evaluating outputs for accuracy and bias.
3. Practise with the sample questions. Each guide includes illustrative questions that show the style and difficulty. They are best-answer scenario questions, not recall, so use them to calibrate rather than to memorise.

Free training on Anthropic Academy

Anthropic’s Academy hosts self-paced courses that map onto the exam domains. They are free, need only an email to enrol, and award a completion certificate that is separate from the exam credential. The full catalogue is at anthropic.skilljar.com, and Anthropic’s wider learning hub is at anthropic.com/learn.

Anthropic has not yet published a course-to-certification mapping of its own. The matrix below is our content-based mapping of the Academy catalogue against each certification’s published exam domains. The markers mean: Foundation where a course builds baseline knowledge for that cert’s domains, Advanced where it covers them at production or deep-dive depth, and blank where it does not apply. Study times are the Academy’s own estimates.

Course	CCAO-F	CCDV-F	CCAR-F	CCAR-P	Study time
Claude 101	Foundation
Claude Code 101		Foundation	Foundation		1.5 hrs
Claude Platform 101		Foundation
Introduction to Claude Cowork	Foundation				0.5 hrs
Claude Code in Action		Advanced	Advanced	Foundation	1 hr
AI Fluency: Framework and Foundations	Foundation				1.1 hrs
AI Capabilities and Limitations	Foundation	Foundation	Foundation	Foundation	0.25 hrs
Building with the Claude API		Advanced	Advanced	Foundation	8.1 hrs
Introduction to Model Context Protocol		Advanced	Foundation	Foundation	1 hr
MCP: Advanced Topics		Advanced	Advanced	Foundation	1.1 hrs
Introduction to Agent Skills		Foundation	Foundation
Introduction to Subagents		Foundation	Foundation
Claude with Amazon Bedrock		Advanced	Foundation	Foundation	8 hrs
Claude with Google Cloud’s Vertex AI		Advanced	Foundation	Foundation	8 hrs
AI Fluency for Educators	Foundation				0.4 hrs
AI Fluency for Students	Foundation				0.5 hrs
Teaching AI Fluency	Foundation				0.6 hrs
AI Fluency for Nonprofits	Foundation				0.9 hrs
AI Fluency for Small Businesses	Foundation				0.9 hrs
AI Fluency for Builders	Foundation				1 hr

Two gaps are worth naming. The Associate exam weights governance and responsible use at 15 percent, but no Academy course treats AI governance or risk as its main subject; it is only touched on within the AI Fluency diligence material. The Architect – Professional exam is the least well served by the catalogue: nothing currently covers its governance, safety and risk domain (including GDPR and HIPAA) or its stakeholder communication and lifecycle domain, which is why the cloud and API courses map to it only at Foundation depth. Both of those domains lean on real delivery experience more than on any course.

Supplementary resources beyond the Academy

Not part of the Academy catalogue, and not officially mapped to any exam, but useful practice material for the technical credentials:

Resource	CCAO-F	CCDV-F	CCAR-F	CCAR-P
Claude Code: A Highly Agentic Coding Assistant (DeepLearning.AI)		Foundation	Foundation
MCP: Build Rich-Context AI Apps (DeepLearning.AI)		Foundation	Foundation
Agent Skills with Anthropic (DeepLearning.AI)		Foundation	Foundation
anthropics/courses on GitHub (API fundamentals, prompt engineering, real-world prompting, prompt evaluations, tool use)	Foundation	Foundation	Foundation	Foundation

To reach the exam itself, registration runs through the Anthropic Partner Academy: anthropic-partners.skilljar.com/page/partner-certifications.

The practical logistics

A few things worth knowing before booking any of the four:

• Delivery. Proctored through Pearson VUE (pearsonvue.com/us/en/anthropic.html), online or at a test centre. Closed book, no external tools, no AI assistance.
• Passing. Scaled score of 720 on a 100 to 1,000 scale. Your report shows a pass or fail plus a percentage-correct breakdown by domain.
• Retakes. If you do not pass, the waiting periods increase with each attempt: 14 days after the first, 30 after the second, 90 after the third. You may sit a given exam up to four times in a rolling 12 months, and the fee applies to each attempt. Limits are per exam, so failing one does not block you from a different one.
• Validity and renewal. 12 months. Renew on time with a free, non-proctored assessment on the Academy. Let it lapse and you retake the full exam at full price.
• Registration. You register and check out through the Partner Academy, then create a Pearson VUE account to schedule. The fee shown at checkout reflects any discount that applies to your partner tier.

nThe details in this post are drawn from the official Anthropic exam guides, version 1.0, effective July 2026. Fees, formats, and availability can change; confirm current details on the Anthropic Partner Academy before booking.