Program Overview

This site covers two programmes side by side: Anthropic’s Claude Certification Program, a paid proctored credential, and OpenAI’s Academy pathways and certification initiative, a free learning ecosystem whose certification layer is still expanding. This page describes the Claude programme; the OpenAI side has its own overview, and Claude or OpenAI? compares them head to head.

The Claude Certification Program is Anthropic’s proctored credentialing track. It launched on 12 March 2026 with a single exam (Architect – Foundations) and expanded in July 2026 to four live credentials delivered through Pearson VUE. All four share the same exam machinery and differ only in blueprint: which skills are tested and how heavily each is weighted.

The three roles, four credentials

Role	Credential	Code	Items	Fee	Validates that you can…
Associate	Associate – Foundations	CCAO-F	60	$99	Apply Claude to business and productivity tasks with minimal guidance and sound judgment
Developer	Developer – Foundations	CCDV-F	53	$125	Build, integrate and ship production-grade applications, agents and workflows
Architect	Architect – Foundations	CCAR-F	60	$125	Design Claude-based systems: agentic architecture, tools/MCP, Claude Code, prompting, context
Architect	Architect – Professional	CCAR-P	63	$175	Own the full solution lifecycle including RAG integration, evaluation, governance and stakeholder communication

What all exams share

• Format: multiple-choice and multiple-response. Every item tells you how many responses to select. There is no penalty for guessing.
• Scoring: scaled 100–1000, pass at 720. The cut score was set by a formal standard-setting study; you pass by meeting a fixed standard, not by beating other candidates.
• Score report: pass/fail, scaled score and percent-correct per domain. Use it to target a retake if needed.
• Style: scenario-first. A situation, constraints, and four (or more) defensible options where only one reflects the correct trade-off. Recall alone will not pass these exams.
• Prerequisites: none. Recommended experience is exactly that – recommended. The credential is awarded on exam performance alone.
• Access gate: registration runs through the Anthropic Partner Academy and requires a work email on a Claude Partner Network domain. Personal addresses are rejected. Joining the Partner Network is free for organisations.

How the courses are built

Each course on this site is built from its official Exam Guide v1.0 blueprint. For every domain you get:

Recommended reading order

1. Read Claude or OpenAI? if you have not chosen a programme yet.
2. Read Which certification? and pick your Claude target.
3. Read Exam logistics so registration does not surprise you.
4. Read Study method and Question anatomy.
5. Work through your course’s domains in weight order – heaviest domain first.
6. Skim the Lexique daily; use the Claude appendix as a reference.
7. Sit the course’s practice exam under timed conditions, review every explanation, then book.

Studying the OpenAI side instead? Start at the OpenAI overview, then the credential landscape and the learning sequence.

A note on currency

This material reflects the model lineup and API surface as of September 2026 (Claude Fable 5.1, Opus 5, Sonnet 5, Haiku 4.5; adaptive thinking; effort levels; tool search; context editing and compaction; Managed Agents; the Claude Agent SDK). The exams are described as model-agnostic, but scenarios assume a current lineup. Where a fact could drift, the appendix tells you how to verify it live.