explainx / blog
Anthropic Academy’s Claude Certified Architect exam is a ~301-level, 120-minute proctored test with 60 multiple-choice questions across agent architecture, MCP, Claude Code, prompting, and reliability. Exam guide, competency breakdown, scenarios, pricing—and how explainx.ai is building prep alongside Claude for Work.
Apr 14, 2026
A ~110-minute Claude productivity curriculum (Projects, Artifacts, Memory, Research Mode, prompting, Claude Code, MCP) now has a public student hub: 20 copy-paste prompts plus curated Anthropic and MCP links.
Jul 1, 2026
Anthropic''s official "Redeploying Fable 5" post — export controls lifted June 30, global restore July 1, stricter classifiers, 50% of weekly limits through July 7, credits after. HN breakdown: coding confusion, Opus fallback billing, invisible competitive-use limits, and Commerce reevaluation risk.
Jul 1, 2026
Fable 5 is live July 1. ClaudeDevs reset 5-hour and weekly limits on restore. Included Fable through July 7 at 50% weekly cap, then credits. Full billing guide.
Anthropic’s Anthropic Academy path now includes a formal milestone for partner practitioners: Claude Certified Architect—a credential aimed at people who already ship with Claude and want vendor-recognized proof of depth.
This post is a field guide: format, domains, scenario framing, pricing, official prep, and how explainx.ai fits in—including timed mock tests and a certification study guide, plus our Claude for Work course on Udemy.
Disclaimer: Exam structure and policies belong to Anthropic and Skilljar (Anthropic’s learning platform). Confirm details on the live Academy pages before you register.
Public copy positions the exam as exclusive to Anthropic partners: practitioners at partner companies with foundational knowledge who are ready to show deeper, applied skill—not an entry-level literacy quiz.
Non-partners see messaging to apply to the Claude Partner Network if they want access.
| Attribute | Detail |
|---|---|
| Duration | 120 minutes, one continuous session |
| Questions | 60 multiple-choice |
| Proctoring | Yes—no external resources, no breaks in-session |
| Results | Score report within ~2 business days with section breakdowns; digital certificate framed as recognizable by organizations using Claude |
| Pricing | $99 per attempt (partner discounts described on the Academy site) |
Treat the table as a summary of Anthropic’s public marketing; seat availability, retake rules, and regional policies can change—check the registration flow.
Anthropic weights 60 questions across five domains:
| Area | Weight | What it emphasizes |
|---|---|---|
| Agentic Architecture & Orchestration | 27% | Agentic loops, coordinator–subagent patterns, task decomposition, session state, workflow enforcement |
| Tool Design & MCP Integration | 18% | Tool boundaries, structured errors, MCP servers, distributing tools across agents |
| Claude Code Configuration & Workflows | 20% | CLAUDE.md hierarchies, slash commands, path-specific rules, plan mode vs direct execution, CI/CD |
| Prompt Engineering & Structured Output | 20% | Explicit criteria, few-shot patterns, JSON-schema enforcement, validation and retry loops |
| Context Management & Reliability | 15% | Long-horizon context, escalation, error propagation in multi-agent setups, uncertainty handling |
If you are mapping study time, agent orchestration plus Claude Code plus structured prompting account for most of the blueprint—MCP and reliability are smaller slices but still fair game in scenario-based items.
The Academy describes six production-style scenarios; each exam randomly draws four. They anchor questions in realistic contexts—for example:
Customer support resolution agent — Claude Agent SDK, high-ambiguity tickets, MCP tools (get_customer, lookup_order, process_refund, escalate_to_human), targets like 80%+ first-contact resolution with safe escalation. (Touches: orchestration, MCP, reliability.)
Code generation with Claude Code — acceleration across generation, refactor, debug, docs; slash commands, CLAUDE.md, plan mode vs execution. (Claude Code, reliability.)
Multi-agent research system — coordinator delegates to web search, document analysis, synthesis, reporting; cited outputs. (Orchestration, MCP, reliability.)
Developer productivity — Agent SDK, codebase exploration, boilerplate, automation; built-in tools plus MCP. (Tools, Claude Code, orchestration.)
Claude Code in CI/CD — automated review, tests, PR feedback; prompts that reduce false positives. (Claude Code, structured output.)
Structured data extraction — unstructured documents → validated JSON, schema enforcement, edge cases, downstream integration. (Structured output, reliability.)
Full narrative detail lives in the official Exam Guide (see below).
Practice exam
3 timed mock exams with shuffled questions, instant scoring, and per-question explanations. Pass score: 720/1000. The fastest way to find your weak domains before exam day.
Anthropic Academy (Skilljar) — Anthropic hosts Academy courses and certification flows on Skilljar (login, enrollments, and FAQs explain why you see Skilljar branding).
Exam access request — Partner-aligned registration is surfaced through Anthropic’s Skilljar paths (for example an access-request route such as anthropic.skilljar.com/claude-certified-architect-foundations-access-request); confirm the exact URL on the live Academy page before you pay or schedule.
Exam guide PDF (Foundations) — full scenarios and official framing as published for the certification track. If Anthropic updates the file, use the download linked from Academy so your copy matches the latest revision.
Recommended Anthropic learning — Academy curricula include courses such as Building with the Claude API, Introduction to Model Context Protocol, Claude Code in Action, and Claude 101—aligned foundations, not a substitute for hands-on building.
Start mock tests → — eight timed practice exams (Foundations drills, domain focus, scenario marathon, 60-question / 120-minute simulation), 1,000+ shuffled multiple-choice questions, instant explanations, $5 lifetime access.
Certification study guide → — domain weights, task statements per competency area, six scenario narratives, in-scope / out-of-scope topics, and technology checklist mapped to the official Foundations Exam Guide.
Our Claude for Work course on Udemy is built for teams and individuals who want hands-on fluency with Claude in real work: productivity workflows, Claude Code, MCP, and practical prompting—overlapping several Architect exam themes (especially Claude Code workflows, tool/MCP literacy, and structured prompting at a practitioner level).
If you are not yet in Anthropic’s partner exam cohort, Claude for Work is still one of the strongest general stacks we publish for “Claude at work” credibility—and it complements Anthropic’s own Academy catalog.
SKILL.md, progressive disclosure, agent workflowsClaude Certified Architect is Anthropic’s bid to standardize what “production-ready” means for Claude builders inside the partner ecosystem: long, closed-book, multi-domain, scenario-grounded. Start from the official Exam Guide and Academy courses, run explainx mock tests for timed MCQ practice, and use Claude for Work on Udemy for applied breadth.
Practice exam
3 timed mock exams with shuffled questions, instant scoring, and per-question explanations. Pass score: 720/1000. The fastest way to find your weak domains before exam day.
Exam names, weights, and policies are summarized from Anthropic’s public Academy messaging; verify on Anthropic Academy before registering. explainx.ai is not affiliated with Anthropic’s certification program.