100+ Free AI-200 Practice Questions

Pass your Microsoft Certified: Azure AI Cloud Developer Associate (AI-200) exam on the first try — instant access, no signup required.

✓ No registration✓ No credit card✓ No hidden fees✓ Start practicing immediately

100+ Questions

100% Free

1 / 10

Question 1

Score: 0/0

Which Azure compute service is the most appropriate choice for hosting a containerized AI inference microservice that needs autoscaling, scale-to-zero, and managed networking, without managing Kubernetes directly?

Azure Container Apps

Azure Service Fabric

Azure Spring Apps

Azure Batch

to track

2026 Statistics

Key Facts: AI-200 Exam

700/1000

Passing Score

Microsoft (typical for role-based associate)

40-60 Q

Typical Questions

Microsoft (typical for role-based associate)

100 min

Exam Duration

Microsoft (typical for role-based associate)

$165

US Exam Fee

Microsoft (typical for role-based associate)

6 domains

Inferred Skills Areas

AI-200T00-A course modules

Annual

Free Renewal

Microsoft Learn

AI-200 is a brand new 2026 Microsoft associate certification for backend and AI cloud developers. Expect roughly 40-60 questions in 100 minutes and a 700/1000 passing score. Inferred domains from the AI-200T00-A course: containerized compute (Container Apps, AKS, Functions), AI-enabled data services (Cosmos DB for NoSQL, PostgreSQL+pgvector, Azure Managed Redis), event-driven architectures (Service Bus, Event Grid, Event Hubs), AI service integration (Azure OpenAI, AI Search, Document Intelligence), identity and secrets (Entra ID, managed identity, Key Vault), and observability (Application Insights, OpenTelemetry).

Sample AI-200 Practice Questions

Try these sample questions to test your AI-200 exam readiness. Each question includes a detailed explanation. Start the interactive quiz above for the full 100+ question experience with AI tutoring.

1Which Azure compute service is the most appropriate choice for hosting a containerized AI inference microservice that needs autoscaling, scale-to-zero, and managed networking, without managing Kubernetes directly?

A.Azure Container Apps

B.Azure Service Fabric

C.Azure Spring Apps

D.Azure Batch

Explanation: Azure Container Apps is a managed serverless container platform built on Kubernetes/KEDA that supports HTTP/event-driven autoscaling, scale-to-zero, revisions, traffic splitting, managed identities, and Dapr — all without exposing the Kubernetes API. It is the recommended starting point for containerized AI microservices.

2An AI Cloud Developer needs to deploy an AI inference container that requires fine-grained control over node pools, GPUs, and Kubernetes resources. Which Azure service should they choose?

A.Azure Functions Consumption plan

B.Azure Kubernetes Service (AKS)

C.Azure App Service Web Apps

D.Azure Logic Apps

Explanation: Azure Kubernetes Service (AKS) gives you full Kubernetes control, including custom node pools, GPU SKUs, and advanced networking. It is appropriate when Container Apps' abstractions are insufficient. It comes with higher operational responsibility than Container Apps.

3Which Azure Functions hosting plan is best for a low-latency AI API endpoint that must avoid cold-start delays?

A.Consumption plan

B.Premium plan or Flex Consumption with always-ready instances

C.Free tier

D.App Service Free F1

Explanation: Premium and Flex Consumption plans for Azure Functions support pre-warmed (always-ready) instances, eliminating cold starts and providing VNet integration. The basic Consumption plan has cold-start latency that can be problematic for user-facing AI APIs.

4An AI cloud app must call Azure OpenAI from Azure Container Apps without storing API keys. Which authentication mechanism should you configure?

A.API key in plain environment variable

B.User-assigned managed identity with Cognitive Services OpenAI User role on the Azure OpenAI resource

C.Personal access token of a developer

D.Anonymous access

Explanation: Azure Container Apps support both system- and user-assigned managed identities. Assigning a user-assigned identity and granting it the Cognitive Services OpenAI User RBAC role on the Azure OpenAI resource enables keyless, RBAC-based authentication without secret management.

5Which Azure data service supports vector embeddings via the pgvector extension and is recommended in the AI-200 curriculum for vector-enabled AI workloads on PostgreSQL?

A.Azure SQL Managed Instance

B.Azure Database for PostgreSQL with pgvector extension

C.Azure Files Premium

D.Azure Table Storage

Explanation: The AI-200 curriculum highlights Azure Database for PostgreSQL with the pgvector extension for storing and querying vector embeddings alongside relational data. pgvector adds a vector type and ANN indexes (HNSW, IVFFlat) suitable for RAG and similarity search.

6Which Azure Cosmos DB API is most appropriate for storing JSON documents with embedded vector embeddings used by an AI app?

A.Azure Cosmos DB for NoSQL with vector search support

B.Cosmos DB for Cassandra

C.Cosmos DB for Gremlin

D.Cosmos DB for Table

Explanation: Azure Cosmos DB for NoSQL supports vector search natively, allowing you to store embeddings alongside JSON documents and query them with vector indexes. It is the AI-200 curriculum's reference NoSQL choice for vector-enabled data patterns.

7Which Azure cache service is positioned in the AI-200 curriculum for caching, streaming, and vector search at low latency?

A.Azure Cache for Memcached

B.Azure Managed Redis (with vector search support)

C.Azure CDN

D.Azure Files

Explanation: Azure Managed Redis (the next-generation managed Redis offering) supports caching, streams, and vector search with Redis's vector index types. The AI-200 curriculum calls it out as the preferred Redis option for AI app caching, streaming, and vector workloads.

8Which Azure messaging service is best for decoupling a long-running AI processing pipeline from a public REST API by buffering work as messages?

A.Azure Service Bus queue

B.Azure DNS

C.Azure ExpressRoute

D.Azure Front Door

Explanation: Azure Service Bus queues provide durable, ordered, transactional messaging suitable for decoupling synchronous APIs from long-running AI processing. Workers consume messages and scale based on queue depth. Service Bus also offers topics for publish/subscribe patterns.

9Which Azure service is best suited as the eventing backbone that triggers an AI processing function when a new file lands in Blob Storage?

A.Azure Event Grid (with Blob Storage events)

B.Azure Notification Hubs

C.Azure Front Door

D.Azure Bastion

Explanation: Azure Event Grid is the platform's general-purpose eventing service. Subscribing to Blob Storage 'BlobCreated' events and routing them to an Azure Function or Container App is the canonical pattern for triggering AI processing on file uploads.

10Which Azure messaging service is designed for ingesting millions of events per second (telemetry, click streams, device data) for downstream AI analysis?

A.Azure Event Hubs

B.Azure Service Bus topics

C.Azure Queue Storage only

D.Azure DNS

Explanation: Azure Event Hubs is the high-throughput event-streaming platform optimized for big-data scenarios such as telemetry and click-stream ingestion. It supports Kafka clients and partitioned consumer groups, making it ideal for feeding AI/ML pipelines at scale.

About the AI-200 Exam

The AI-200 exam validates the skills needed to build, integrate, and operate AI-driven cloud applications on Azure using containerized compute, vector-enabled data services, event-driven AI pipelines, serverless functions, identity-aware application design, secrets management, and distributed observability.

Questions

50 scored questions

Time Limit

100 minutes

Passing Score

700/1000

Exam Fee

$165 USD (Microsoft / Pearson VUE)

AI-200 Exam Content Outline

20-25%

Compute and Containerization Patterns

Host AI workloads with Azure Container Apps, Azure Kubernetes Service (AKS), and Azure Functions. Configure revisions, traffic splits, scale rules (KEDA), GPU node pools, jobs, sidecars (Dapr), and serverless triggers.

20-25%

AI-Enabled Data Services

Design and query data with Azure Cosmos DB for NoSQL (vector search, change feed, integrated cache), Azure Database for PostgreSQL with pgvector (HNSW/IVFFlat), Azure Managed Redis (caching, streams, vector search), Blob Storage, and Azure AI Search.

15-20%

Event-Driven and Message-Based Architectures

Integrate services using Azure Service Bus (queues, topics, DLQ), Azure Event Grid (CloudEvents pub/sub), Azure Event Hubs (high-throughput streaming, Kafka endpoint), Azure Notification Hubs, Logic Apps, and Durable Functions.

15-20%

AI Service Integration and Resilience

Call Azure OpenAI (chat completions, function calling, structured outputs, streaming), Azure AI Search (hybrid + semantic), Document Intelligence, Speech, and Language. Implement retries, exponential backoff with jitter, circuit breakers, semantic caching, and APIM-based AI gateway patterns.

10-15%

Identity, Security, and Configuration

Use Microsoft Entra ID with managed identity (system / user-assigned, workload identity, federation), Azure Key Vault (secret rotation), App Configuration (feature flags), App Service Authentication, private endpoints, VNets, NSGs, WAF, and Defender for Containers.

10-15%

Distributed Observability and CI/CD

Instrument with OpenTelemetry into Application Insights (Application Map, distributed traces, release annotations), centralize logs in Log Analytics with KQL, build dashboards in Workbooks, and deploy with Bicep + GitHub Actions (OIDC, environment protection, slots, azd templates).

How to Pass the AI-200 Exam

What You Need to Know

Passing score: 700/1000
Exam length: 50 questions
Time limit: 100 minutes
Exam fee: $165 USD

Keys to Passing

Complete 500+ practice questions
Score 80%+ consistently before scheduling
Focus on highest-weighted sections
Use our AI tutor for tough concepts

AI-200 Study Tips from Top Performers

1Build at least one end-to-end Azure AI cloud app: Container App + Cosmos DB for NoSQL + Azure OpenAI + Azure AI Search + managed identity + Application Insights, deployed via Bicep and GitHub Actions.

2Practice both vector-enabled patterns the curriculum highlights: pgvector on Azure Database for PostgreSQL and vector search on Cosmos DB for NoSQL. Know when each is preferable.

3Master the Azure Container Apps mental model: environments, revisions, traffic splits, KEDA scalers, jobs, sidecars/Dapr, internal vs external ingress.

4Be fluent with the Azure messaging trio: Service Bus (transactional, ordered), Event Grid (pub/sub, CloudEvents), Event Hubs (high-throughput streaming, Kafka endpoint).

5Memorize keyless authentication patterns: managed identity + RBAC, OIDC federation for GitHub Actions, workload identity for Container Apps and AKS.

6Practice OpenTelemetry instrumentation feeding Application Insights — including W3C trace context propagation, Application Map, and release annotations.

7Know APIM's Azure OpenAI policies (token-based throttling, semantic caching, load balancing), since APIM is the recommended AI gateway.

Frequently Asked Questions

What is the AI-200 exam?

AI-200 is the Microsoft exam for the Azure AI Cloud Developer Associate credential introduced in 2026. It validates the skills needed to build, integrate, monitor, and operate AI-driven cloud applications on Azure across containerized compute, AI-enabled data services, event-driven architectures, identity, secrets, and distributed observability.

How is AI-200 different from AZ-204?

AI-200 replaces the AZ-204T00 training course, which retires May 29, 2026, and is purpose-built around AI workloads. While AZ-204 emphasized general Azure development on App Service, Functions, and storage, AI-200 centers on containerized compute, vector-enabled data services (Cosmos DB for NoSQL, PostgreSQL+pgvector, Azure Managed Redis), event-driven AI pipelines, and integrating Azure OpenAI and Azure AI services with proper identity, security, and observability.

Is AI-200 still in beta?

Yes — at the time of writing, AI-200 is brand new for 2026. The AI-200T00-A training course was scheduled to publish on April 30, 2026, with the exam expected to enter beta around the same time and to go live around July 2026. Microsoft had not yet published the formal Skills Measured outline; the domains shown here are inferred from the AI-200T00-A course modules and the official AI-200 product description. Always check the official Microsoft Learn page for the latest beta or live status before scheduling.

How many questions are on AI-200 and how long do you get?

Microsoft role-based associate exams typically deliver about 40-60 questions. For AI-200, plan for a 100-minute exam duration and a 700 out of 1000 passing score, with possible interactive case-study and scenario items as Microsoft typically uses on associate exams.

How long should I study for AI-200?

Plan for about 80-140 hours over 6-10 weeks depending on prior Azure development experience. Effective preparation includes building at least one production-style AI app: a Container App or Function backend that calls Azure OpenAI with managed identity, persists state to Cosmos DB or PostgreSQL+pgvector, processes events via Service Bus or Event Grid, and is monitored end-to-end via Application Insights. Then complete 100+ practice questions.

Does AI-200 certification expire?

Yes. Microsoft associate certifications expire 12 months after you earn them. You can renew at no cost by passing a free online renewal assessment on Microsoft Learn before the expiration date.