The short version
Take AIF-C01 if you need to understand and discuss AI on AWS. Take MLA-C02 if you build, deploy, and run ML or generative AI systems and already have about a year of hands-on AWS experience. If you are unsure, start with AIF-C01: it is foundational, and AWS does not require it for the engineer exam. Take both only if your role moves from understanding to building. MLA-C02 is listed as a beta as of October 22, 2026, so some of its details can still change.
Side by side
| AIF-C01 | MLA-C02 | |
|---|---|---|
| Credential | AWS AI Practitioner | AWS Machine Learning Engineer – Associate |
| Level (AWS) | Foundational | Associate |
| Audience | People who use but do not necessarily build AI/ML on AWS | ML engineers, MLOps and LLMOps engineers, data engineers, backend developers, data scientists |
| Experience AWS describes | Up to 6 months of exposure to AI/ML on AWS | At least 1 year with Amazon SageMaker AI, Amazon Bedrock, and other ML engineering services |
| Focus | Concepts, choosing the right approach, responsible use | Building, deploying, orchestrating, monitoring, and securing |
| Domains | 5 domains: 20%, 24%, 28%, 14%, 14% | 4 domains: 28%, 24%, 24%, 24% |
| Question types | Multiple choice, multiple response, ordering, matching | Multiple choice, multiple response |
| Format | 65 questions (50 scored), 90 minutes | Beta page: 85 questions, 170 minutes; exam guide: 50 scored plus 15 unscored |
| Passing score | 700 of 1,000 | 720 of 1,000 (exam guide; AWS says pass/fail does not apply to the beta) |
| Validity | 3 years | 3 years |
| Renewal | Latest AIF-C01, or the Machine Learning Engineer or Generative AI Developer exam | Latest MLA exam or the Generative AI Developer – Professional exam; or maintain for one year in AWS Skill Builder |
| Status | Current, guide version 1.2 (October 29, 2026) | Beta; an AWS blog from September 2026 gives January 14, 2027 for the start of general-availability delivery (the certification page still said TBD on October 27, 2026) |
Format numbers for MLA-C02 come from two AWS sources that do not match. Read the MLA-C02 exam format page and confirm on the AWS page when you book. The MLA-C02 renewal page explains the renewal and maintenance options.
What each exam is really about
AIF-C01 asks “which tool and why”. A question describes a business situation, and you pick the service, technique, or setting that fits, or spot the risk. AWS lists building models, tuning hyperparameters, and building pipelines as outside its scope.
MLA-C02 asks “how do you make it work and keep it working”. Its tasks include ingesting and preparing data, training and tuning models, choosing endpoints and scaling, setting up CI/CD pipelines, building agentic workflows, monitoring, and securing the system. The guide also lists what is out of scope: designing full end-to-end AI solutions and setting ML strategy.
Where they overlap
- Generative AI vocabulary: foundation models, embeddings, retrieval augmented generation, and Amazon Bedrock appear in both.
- Core AWS services: IAM, encryption, and the shared responsibility model matter in each.
- Evaluating models: metrics and trade-offs show up in both, at very different depths.
- Security and responsible use: AIF-C01 tests them as concepts, MLA-C02 as things you configure.
Where they differ
- Depth. AIF-C01 wants you to explain an idea. MLA-C02 wants you to pick the right instance type, endpoint, pipeline step, or monitoring tool.
- Breadth of AI. AIF-C01 includes responsible AI and governance as two full domains (28% together). MLA-C02 focuses on the engineering lifecycle.
- Experience. AWS recommends at least a year of hands-on ML engineering for MLA-C02. AIF-C01 expects up to six months of exposure.
- Length. The beta MLA-C02 listing shows nearly twice the time of AIF-C01.
How to choose
- You manage projects, write requirements, sell, or advise: AIF-C01.
- You are a developer or IT person new to AI: AIF-C01 first as a map, then decide.
- You deploy models or run ML pipelines now: MLA-C02, and AIF-C01 is optional.
- You want the stronger credential and already work with SageMaker AI and Bedrock: MLA-C02. AWS also lists passing it as a way to recertify the AI Practitioner credential.
- You want to start small and see if AI on AWS suits you: AIF-C01.
Order matters little for the exams themselves, since AWS names no prerequisite for either. The knowledge does build, though: AIF-C01 gives you the vocabulary that MLA-C02 assumes.
What about the beta?
Because MLA-C02 is a beta, the English-only beta may change before launch. An AWS blog from September 2026 gives January 14, 2027 as the start of general-availability delivery, while the certification page still said TBD on October 27, 2026. If you hold older MLA-C01 material, AWS’s comparison page says the new version adds topics such as RAG, embeddings, vector stores, and agents. Check the MLA-C02 exam guide and the AIF-C01 exam guide before you commit.
More guides for AIF-C01
- AIF-C01 Exam Format, Scoring and Retakes
- AIF-C01 Study Plan: 4-Week and 8-Week
- AIF-C01 Domains Explained With Weightings
- AIF-C01 Sample Questions With Answers
- Is the AIF-C01 Exam Hard?
- AIF-C01 Prerequisites and Experience
- AIF-C01 Renewal and Retakes Explained
- AIF-C01 Cheat Sheet: Key Facts to Review
- AWS AIF-C01 vs AI-901: Which First?
Next step
Read the AIF-C01 overview and the MLA-C02 overview side by side. Our practice sets are here: AIF-C01 practice tests and MLA-C02 practice tests. For the first exam, see also is AIF-C01 worth it.
Frequently asked questions
Do I need AIF-C01 before MLA-C02?
No. AWS lists no required prerequisite for MLA-C02. The engineer exam does assume hands-on experience.
Does passing MLA-C02 renew AIF-C01?
AWS’s certification and recertification pages say passing the Machine Learning Engineer – Associate exam recertifies the AI Practitioner credential.
Which one is harder?
MLA-C02 is a higher level, longer, and expects practical experience. Whether it feels harder depends on your background.