On December 7, 2026, the Claude Certified Associate - Foundations (CCAO-F) exam changes shape. It moves from seven domains to six, every domain gets a new name or a new weight, the skill list is rewritten around real work tasks, and the exam grows from 60 to 70 questions (60 scored and 10 unscored) in the same 120 minutes. Anthropic published the updated exam guide (version 1.1, effective December 7) on November 17, 2026. This page lists every change and helps you decide which version to take.
The short version
- Format: 70 questions (60 scored, 10 unscored) in 120 minutes, up from 60. Price ($99) and passing score (720) stay the same.
- Domains: seven become six. Troubleshooting and Optimization (10%) is no longer a domain, and the other six are renamed and reweighted.
- Skills: the current guide's 30 one-line objectives become 21 skills, each described as a task, such as diagnosing a weak prompt or deciding how much review an output needs.
- New topics: Skills, connectors, memory and incognito chats, effort levels, code execution for exact numbers, and Anthropic's Delegation and Diligence ideas.
- Before December 7 you sit the current exam. On or after it, prepare from the updated guide.
The format: 70 questions in 120 minutes
CCAO-F before and after December 7, 2026
| Until December 6, 2026 | From December 7, 2026 | |
|---|---|---|
| Exam guide | Version 1.0 (July 2026) | Version 1.1 (effective December 7, 2026) |
| Questions | 60, all scored | 70: 60 scored, 10 unscored |
| Time | 120 minutes | 120 minutes |
| Domains | 7 | 6 |
| Item format | Multiple-choice and multiple-response; each item states how many responses to select | Scenario-based multiple-choice; each item states how many responses to select |
| Passing score | 720 (scaled 100 to 1,000) | 720 (scaled 100 to 1,000) |
| Price | $99 | $99 |
| Languages | English and Japanese | English and Japanese |
The scored count stays at 60, but you now answer 70 questions in the same time, about 1.7 minutes each instead of 2. Ten are unscored trial items that you cannot tell apart, so treat every question as if it counts. Every item is scenario-based: a work situation, then a judgment about what to do with Claude.
Try 15 CCAO-F practice questions free
They come from the CCAO-F practice tests, and each option shows its explanation once you answer.
- 6 timed practice exams
- Every answer explained
- A score for each domain
- $19.99 one-time
Seven domains become six
CCAO-F domains, before and after
| Until December 6 | Weight | From December 7 | Weight |
|---|---|---|---|
| Output Evaluation and Validation | 21% | Output Evaluation and Quality Control | 20% |
| Workflow Integration and Solution Design | 16% | Workflow Design and Integration | 18% |
| Governance, Risk, and Responsible Use | 15% | Responsible Use, Data Protection, and Governance | 17% |
| Prompting and Task Execution | 14% | Task Definition and Prompting | 15% |
| Product and Model Selection | 12% | Capability, Model, and Context Selection | 15% |
| Configuration and Knowledge Management | 12% | Configuration and Knowledge Management | 15% |
| Troubleshooting and Optimization | 10% | No longer a separate domain |
The 10% that Troubleshooting and Optimization carried is spread across the other six, and its ideas live on inside them: diagnosing and iterating on a weak prompt is now a Task Definition and Prompting skill, and working out why an established setup has degraded is now a Configuration and Knowledge Management skill. The biggest gains go to selection (12% to 15%, now covering capabilities and context as well as models) and configuration (12% to 15%).
Every domain's new skills
Task Definition and Prompting (15%)
- Prompt Construction: role, context, task, output format and the context Claude does not have by default, so the first response is fit for purpose.
- Task Decomposition and Sequencing: staged prompts, follow-ups, Project workflows and validation checkpoints for deliverables too big for one request.
- Prompt Diagnosis and Iteration: trace a weak output to a missing context, an ambiguous instruction, an absent constraint or the wrong strategy, and decide whether to iterate, restart or change approach.
Output Evaluation and Quality Control (20%)
- Output Evaluation and Fitness for Purpose: judge accuracy, completeness and relevance, and classify an output as ready to use, needing revision or needing human correction.
- Hallucination Detection and Verification: spot fabricated citations, confident invented specifics, inconsistencies and bias, and verify before use.
- Review Depth and Escalation: set how much review an AI-assisted product needs from the consequence of an error, internal draft versus external delivery, and what must be escalated.
- Output Adaptation and Format Selection: adapt an output to its audience and channel, including artifacts, structured data, code-executed results for exact computation and Skill-produced documents.
Capability, Model, and Context Selection (15%)
- Capability Selection: pick the right mix of chat, Projects, Artifacts, research, code execution, Skills, connectors and memory for a task.
- Model and Effort Selection: match model tier and effort level to quality, speed and cost, including when the faster, cheaper option is the right one.
- Context and Memory Management: decide when to restart, summarize, persist to a Project or rely on memory, and keep memory clean with reviews, edits, deletions and incognito chats.
Workflow Design and Integration (18%)
- End-to-End Deliverable Workflows: run a research, analysis or planning deliverable from requirements through cited research and code-executed analysis to an actionable plan.
- Delegation and Workflow Redesign: assign each step of a workflow to Claude, a person or a person reviewing Claude, using reversibility, stakes, the need for human judgment and accountability, and recognize over-delegation.
- Repeatable Workflow Design: turn a recurring manual task into a consistent workflow with Projects, Skills, connectors and code-executed steps that a colleague can own.
- Capability Communication: explain to stakeholders what Claude can and cannot reliably do for a use case, without overstating or understating it.
Configuration and Knowledge Management (15%)
- Project Configuration: instructions, knowledge sources, Skills and scoped memory for a recurring use case, and which of the three each requirement belongs in.
- Connector and Knowledge Management: connecting sources such as Google Drive and Gmail, uploads versus connectors, and keeping knowledge current and free of noise.
- Persistent-Instruction Design: what belongs in persistent instructions versus each prompt, so repeated work stays consistent.
- Configuration Lifecycle and Degradation: find which layer (instructions, stale knowledge, Skill versions or accumulated memory) caused a quality drop, and fix it.
Responsible Use, Data Protection, and Governance (17%)
- Use-Case Appropriateness: classify a use case as appropriate, appropriate with safeguards or inappropriate, considering delegation limits, trust in third-party Skills and connectors, and the Anthropic Usage Policy.
- Data Protection in Practice: classification, redaction and anonymization, judgment about uploads for code-executed analysis, and memory and incognito controls.
- Policy, Ethics, and Diligence: compare usage against policy, spot ethical risks and Diligence lapses, and tell a policy gap from a training gap.
What is new for CCAO-F candidates
The current guide talks about prompting, evaluation and Projects in general terms. The new one names product features and judgment calls. Expect questions about Skills (including third-party ones), connectors, memory and incognito chats, effort levels, code execution when numbers must be exact, and artifacts. Expect Anthropic's AI Fluency vocabulary, such as Delegation and Diligence, to frame workflow and governance questions. And expect more questions about how much checking an output needs, based on what happens if it is wrong. The guide still recommends regular, hands-on use of Claude at work as the best preparation, with no prerequisite.
Should you test before or after December 7?
- Nearly ready on the current guide? Book a date before December 7 and sit the seven-domain version you studied for.
- Starting now? Study from the updated guide and book on or after December 7. Spend time in Claude itself with Projects, Skills, connectors and memory, because the new skills assume you have used them.
- Not sure? The two versions share most of their ground: prompting, evaluating outputs, choosing models, Projects and responsible use. If your study so far covered those, the move to the new version is mostly about the newer features and the review-depth judgments.
How to prepare for the new version
- Read the updated exam guide, including its revised sample questions.
- For each skill above, do the task once in Claude: set up a Project with a Skill and a connector, review memory, run a code-executed analysis, and decide the review level for one real document.
- Use the CCAO-F complete guide and the exam domains breakdown, which are being revised to the new guide.
- Time yourself with the CCAO-F practice tests. Six follow the current guide for exams until December 6, and 70-question practice tests on the December 7 blueprint are being added next to them; the first went live on November 19, 2026.
Frequently asked questions
Sources: