AI-200

AI-200: Azure AI Cloud Developer

Everything you need for Microsoft exam AI-200: format, passing score, the four skill areas, a four-week study plan and free sample questions.

Duration
120 min
Passing score
700 / 1000
Replaces
AZ-204
Languages
13

AI-200, Developing AI Cloud Solutions on Azure, is Microsoft’s exam for developers who build the back end of AI applications: the containers they run in, the databases that hold their embeddings, the messaging that connects them, and the monitoring that keeps them honest. Passing it earns the Microsoft Certified: Azure AI Cloud Developer Associate credential.

It is the successor to AZ-204. The Azure Developer Associate certification and its exam were retired on July 31, 2026, and AI-200 is now the associate exam for Azure developers.

Do not mistake it for a model exam. You will not be asked to tune a prompt or choose a language model. AI-200 tests whether you can build the infrastructure an AI app depends on — vector search in Cosmos DB and PostgreSQL, Redis caching, Service Bus and Event Grid, Functions, Container Apps and AKS — with Python as the named language.

What the exam covers

Skill areaWeight
Develop containerized solutions on Azure20–25%
Develop AI solutions by using Azure data management services25–30%
Connect to and consume Azure services20–25%
Secure, monitor, and troubleshoot Azure solutions20–25%

The data domain is the largest and the newest. Vector similarity search appears in three different services, so expect to be asked which one fits a scenario, not just how each works.

Where to start

Read the exam format for the mechanics, then what changed from AZ-204 if you studied for the old exam. The study plan fits in four weeks if you already write code against Azure.

AI-200 guides

  1. AI-200 exam format: questions, score and costThe AI-200 format: 120 minutes, passing score 700, 13 languages, question types, pricing and what the four domain weights mean for your revision.
  2. AI-200 vs AZ-204: what changed?AZ-204 retired on 31 July 2026 and AI-200 replaced it. What carries over, what was dropped, and the vector database content that is entirely new.
  3. AI-200 study plan: pass in 4 weeksA four-week AI-200 study plan: containers, vector data services, messaging and Functions, then security and monitoring, with a lab for every week.
  4. Free AI-200 sample questions with answersFive free AI-200 practice questions on containers, vector search, messaging, Key Vault and KQL, each with an explained answer.
  5. AI-200 containerized solutions explainedThe AI-200 container domain explained: Container Registry and ACR Tasks, App Service, Container Apps revisions and KEDA, and AKS manifests.
  6. AI-200 data management services explainedThe largest AI-200 domain explained: Cosmos DB vector search and change feed, PostgreSQL with pgvector, and Azure Managed Redis caching.
  7. AI-200 connect to Azure services explainedThe AI-200 connect domain explained: Service Bus queues, topics and dead-lettering, Event Grid filters and retries, and Azure Functions bindings.
  8. AI-200 secure, monitor and troubleshootThe AI-200 security and monitoring domain explained: Key Vault secrets and rotation, App Configuration, OpenTelemetry tracing and KQL queries.
  9. Is AI-200 hard? Exam difficulty explainedHow hard AI-200 really is: why the vector database depth, the code snippets and the thin study material matter more than the number of services.
  10. Is AI-200 worth it? Who should take itIs AI-200 worth taking? Who benefits, what it costs, and why the retirement of AZ-204 makes it the associate exam for Azure developers.
  11. Free AI-200 practice test: 20 questionsA free 20-question AI-200 practice test on containers, vector data services, messaging, Functions, Key Vault and KQL, with explained answers.
  12. AI-200 study resourcesWhat to study for AI-200, which AZ-204 material to avoid, and the six labs that cover most of the exam better than reading does.
  13. AI-200 cheat sheetA one-page AI-200 reference: the four skill areas, which Azure service answers each requirement, vector index choices, KQL basics and common traps.