Anthropic's certification program grew from two credentials to four on July 23, 2026, and the expansion did something more useful than add rungs to a ladder: it split the entry level in two. The Claude Certified Associate - Foundations (CCAO-F) certifies professionals who use Claude through the claude.ai application as a daily working tool, meaning people in operations, marketing, project management, education, and communications who draft, analyze, research, and plan with it. The Claude Certified Developer - Foundations (CCDV-F) certifies engineers who build software on top of Claude through the API (the programmatic interface that applications use to send requests to the model), the client SDKs (prebuilt code libraries that wrap the API for languages like Python and TypeScript), and Claude Code, Anthropic's command-line coding agent. Both sit at the Foundations tier, both run under the same proctoring and scoring rules, and both grant a digital badge through Credly by Pearson. What separates them is which side of the product you stand on: CCAO-F measures how well you work with the finished application, while CCDV-F measures how well you build with the platform underneath it.
That split makes the choice unusually clean for most readers, and this guide lays out the evidence: a side-by-side format comparison, a domain-level look at what each exam really asks, the technical-fluency line that decides borderline cases, and a decision framework by role and career goal. For the individual deep dives, see the complete CCAO-F guide and the complete CCDV-F guide.
Quick Verdict
If your work lives in documents, briefs, analyses, and plans, and you reach Claude through the claude.ai app in a browser, take CCAO-F: it costs $99, requires zero coding, and tests the judgment that makes business use of AI reliable. If you write code for a living and your work lives in an editor and a terminal, take CCDV-F: it assumes programming experience and tests the engineering surface of the Claude platform. Either way, calibrate with the domain-weighted CCAO-F practice tests or CCDV-F practice tests before you book a date.
At a Glance
CCAO-F vs CCDV-F at a Glance
| CCAO-F | CCDV-F | |
|---|---|---|
| Full name | Claude Certified Associate - Foundations | Claude Certified Developer - Foundations |
| Track | Business and productivity (claude.ai product) | Developer (API, SDKs, Claude Code, MCP) |
| Announced | July 23, 2026 | July 23, 2026 |
| Questions | 60, all scored | 53, all scored |
| Duration | 120 minutes | 120 minutes |
| Passing score | 720 / 1000 (scaled) | 720 / 1000 (scaled) |
| Cost | $99 USD | $125 USD |
| Domains | 7 | 8 |
| Coding required | None | Python and/or TypeScript, REST APIs, CLI fluency |
| Recommended experience | Regular hands-on professional Claude use | 1-5 years software engineering plus 6+ months with Claude or comparable LLM systems |
| Delivery | Pearson VUE: online proctored or test center | Pearson VUE: online proctored or test center |
| Registration | Anthropic Partner Academy | Anthropic Partner Academy |
| Validity | 1 year, free renewal assessment | 1 year, free renewal assessment |
| Official study source | claude.ai Help Center | docs.anthropic.com |
Because both credentials launched inside the same program on the same day, the mechanics are identical. Each runs 120 minutes through Pearson VUE, either online proctored or at a physical test center, with identity verification required. Registration for both goes through the Anthropic Partner Academy, which requires joining the Claude Partner Network (free at claude.com/partners). Both use scaled scoring on a 100-1000 scale with a 720 passing bar, and the score report shows percent-correct per domain. Both mix single-answer multiple choice with multiple-response items ("Select TWO" or "Select THREE") at roughly a quarter of the exam, and both stay valid for one year with a free, non-proctored renewal assessment on the Partner Academy before the 12-month expiration.
The differences all trace back to audience. CCAO-F asks 60 questions across 7 domains for $99 and assumes only that you use Claude regularly at work. CCDV-F asks 53 questions across 8 domains for $125 and assumes one to five years of software engineering plus at least six months of hands-on work with Claude or comparable large language model (LLM) systems.
Preparing for CCAO-F? Practice with 390+ exam questions
What Each Exam Actually Tests
CCAO-F Tests Judgment Inside the claude.ai Application
CCAO-F is a professional-competence exam with no code anywhere on it. Its seven domains are 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%), and Troubleshooting and Optimization (10%).
The weighting tells you what Anthropic considers the core skill: trust calibration. The heaviest domain is about evaluating what Claude produces, which covers spotting hallucinations (confident but fabricated claims), identifying bias, fact-checking against sources, knowing when a human reviewer must sign off, and refining a draft into something you would put your name on. The governance domain extends that judgment to the organization: which use cases are appropriate, how sensitive data and privacy rules constrain what you paste into a chat, and how company AI policy applies to daily work.
The rest of the blueprint covers the claude.ai product surface. Expect questions on Projects (persistent workspaces that hold instructions and reference documents), Artifacts (standalone outputs like formatted documents that Claude builds beside the conversation), research mode, connectors that link Claude to tools such as Google Drive and Gmail, and choosing among the Haiku, Sonnet, and Opus models for a given task. Scenarios come from business life: a marketing brief that needs three audience variants, meeting notes to turn into an action list, a research summary a stakeholder will challenge. The CCAO-F domain breakdown walks through all seven domains with study guidance.
CCDV-F Tests Engineering on the Claude Platform
CCDV-F is a builder exam. Its eight domains are Applications and Integration (33%), Model Selection and Optimization (17%), Agents and Workflows (15%), Prompt and Context Engineering (11%), Tools and MCPs (10%), Security and Safety (8%), Claude Code (3%), and Eval, Testing, and Debugging (3%).
A third of the exam lives in the heaviest domain, which covers API mechanics: constructing message requests, streaming responses, sending images, enabling extended thinking, prompt caching, choosing batch versus realtime processing, and managing configuration files like CLAUDE.md and settings.json. The agents domain covers the Claude Agent SDK and the judgment call between a workflow (a fixed pipeline of predetermined steps) and an agent (a system where the model decides its own steps and tool calls). The tools domain covers function calling and the Model Context Protocol (MCP), the open standard for connecting Claude to external tools and data sources, including how to build an MCP server. Security questions probe prompt-injection defense, data leakage, and API key management, and the small evaluation domain tests whether you can trace a failure to the integration layer or the model itself. The CCDV-F domain breakdown covers each domain in depth.
Same Theme, Different Question
Several themes appear on both blueprints, and comparing how each exam frames them is the fastest way to feel the gap between the tracks.
Where the Blueprints Mirror Each Other
| Theme | CCAO-F asks about | CCDV-F asks about |
|---|---|---|
| Prompting | Effective business prompts, task decomposition, iterating on a weak draft (14%) | Context-window management, compaction, structured output, defensive parsing (11%) |
| Model selection | Haiku vs Sonnet vs Opus in the product picker; cost, speed, and quality per task (12%) | Tokens, sampling, extended thinking, prompt caching, cost management (17%) |
| Evaluating quality | Hallucination spotting, fact-checking, when human review is required (21%) | Trace analysis, isolating integration failures from model failures (3%) |
| Responsible use | Data sensitivity, org AI policy, appropriate use cases, ethics (15%) | Prompt-injection defense, PII protection, guardrail layering, key management (8%) |
| Configuration | Projects, knowledge sources, connectors like Google Drive and Gmail (12%) | CLAUDE.md hierarchy, settings.json, model pinning, prompt versioning (inside the 33% domain) |
| Automation | Fitting Claude into existing business workflows (16%) | Agent SDK construction, manager and subagent hierarchies (15%) |
Model selection is the clearest illustration. Both exams name Haiku, Sonnet, and Opus and both reward matching the model to the job. CCAO-F frames the question inside the product: which model handles a long research synthesis versus a quick rewrite, and when a long conversation should be restarted or summarized. CCDV-F frames the same choice in engineering units: tokens (the text chunks models process and bill by), latency budgets, and whether prompt caching (reusing a stable prompt prefix across calls at reduced cost) changes the math. One topic, two vocabularies, and the vocabulary is exactly what each exam is testing.
The Technical-Fluency Line
If the role descriptions leave you unsure, one test settles it: can you read a short block of Python or TypeScript and explain what it does, and could you call a REST API (the standard web mechanism where a program sends a structured request to a URL and gets structured data back) with an authentication key? CCDV-F assumes yes without ever asking. Its recommended profile includes one to five years of software engineering, working fluency in Python or TypeScript, and comfort in a command-line interface, and its questions present code-adjacent scenarios such as a streaming response that terminates early or a tool definition an agent keeps misusing.
CCAO-F assumes none of that. Its recommended profile is regular professional Claude use, structured problem-solving, and a practical grasp of AI limitations such as hallucinations, context constraints, and data sensitivity. Those are real skills that plenty of technical people lack, which is why the fluency line runs in both directions: an engineer would find CCAO-F's stakeholder and governance scenarios outside their daily practice, and an analyst would find CCDV-F's API mechanics unanswerable. Sit the exam whose scenarios describe your actual week.
Which Should You Take? A Decision Framework
You Are an Analyst, Project Manager, Marketer, or Operations Professional
Take CCAO-F. The exam tests the exact moves your work already requires: decomposing a fuzzy request into promptable tasks, validating output before it reaches a client or an executive, configuring a Project so Claude carries your team's context, and staying inside data-handling policy while doing all of it. The $99 fee and the absence of any coding requirement make it the lowest-friction credential in the program, and the 21% evaluation domain means passing it says something meaningful about your reliability with AI-assisted work. Start with the how to pass CCAO-F guide and pace your preparation with the 3-week study plan.
You Are a Software Engineer, or Actively Becoming One
Take CCDV-F. The blueprint reads like the checklist for shipping a production Claude feature: API integration, model and cost optimization, agent construction, MCP, security, and debugging. Hiring managers screening for AI engineering skills care about precisely this surface, and the credential also opens the path upward through the architect track. If you are mid-transition into engineering, treat the recommended profile honestly: with less than a year of programming experience, budget extra preparation time for the 33% integration domain, because it rewards habits formed by real API work.
You Sit Between the Tracks
Technical program managers, solutions consultants, and analysts who write occasional scripts face the genuine version of this decision. Apply the fluency line first: if reading code is comfortable, CCDV-F is the stronger signal because it is scarcer among non-engineers and it validates upward into the architect track. If code fluency is partial, CCAO-F today beats a struggling CCDV-F attempt, and nothing stops you from adding the developer credential after deliberate API practice.
Deciding by Career Goal
Choose by where you want to be in two years rather than only by your current title. If the goal is an AI enablement, adoption, or champion role, meaning the person who makes an organization's claude.ai rollout actually work, CCAO-F is the credential that matches the job description, and holding it positions you to lead team-wide certification. If the goal is AI engineering or solution architecture, CCDV-F is the entry point to a track that continues through two more exams, covered below.
Master These Concepts with Practice
Our CCAO-F practice bundle includes:
- 6 full practice exams (390+ questions)
- Detailed explanations for every answer
- Domain-by-domain performance tracking
30-day money-back guarantee
Should One Person Take Both?
For most individual contributors, no. The exams certify different jobs, and a second credential outside your role adds little hiring signal. The exception is bridge roles. AI enablement leads, consultants advising both business teams and engineering teams, and founders or team leads who roll out claude.ai across an organization while also shipping API-backed features get real mileage from the pair, because their credibility depends on fluency with both audiences. The pair costs $224 in first-attempt fees, and free annual renewals mean no recurring cost. Take the one matching your current responsibilities first, so the preparation compounds with your daily work.
Where Each Credential Leads
The two tracks point in different directions after the badge arrives. CCDV-F is the first step on a ladder: the natural next exam is the Claude Certified Architect - Foundations (CCA-F), which shifts from application building toward system design, and the ladder tops out at the Claude Certified Architect - Professional (CCAR-P), the enterprise architecture credential. Our CCDV-F vs CCA-F comparison covers the first handoff in detail.
CCAO-F leads outward rather than upward. It is becoming the org-wide AI-fluency credential: the badge a company can put behind every analyst, marketer, and coordinator to demonstrate that its Claude adoption rests on trained judgment instead of enthusiasm. That makes it valuable to teams in a way the developer track cannot match, since certifying fifty business users is realistic while certifying fifty engineers rarely is. An individual CCAO-F holder who later learns to code has a clean next step waiting in CCDV-F.
Frequently Asked Questions
Choose, Then Calibrate With a Practice Test
The framework gives you the decision, and a timed baseline exam gives you the evidence: a first score tells you whether your date is three weeks away or eight, and which domains need the hours.
Preporato covers both tracks with domain-weighted practice exams. The CCAO-F practice tests include 6 full-length, 60-question exams built on the 7-domain blueprint plus a 500-card flashcard deck, all framed in the business scenarios the real exam uses. The CCDV-F practice tests include 6 full-length, 53-question exams mirroring the 8-domain blueprint with roughly a quarter multiple-response items, plus their own 500-card deck. Both sets come with explanations for every answer, so each miss teaches the concept behind it.
Key Takeaways
0/8 completedSources:
- Anthropic: Four Role-Based Claude Certifications
- Anthropic Partner Academy
- Anthropic Documentation (CCDV-F)
- claude.ai Help Center (CCAO-F)
Last updated: August 8, 2026
Ready to Pass the CCAO-F Exam?
Join thousands who passed with Preporato practice tests
![CCAO-F vs CCDV-F: Which Claude Certification Should You Take? [2026]](/blog/ccao-f-vs-ccdv-f-which-claude-certification-2026.webp)