MLA-C02 vs AIF-C01: which AWS AI exam?

Updated September 20, 2026

AWS has two AI certifications below the specialty tier, and the gap between them is wide. MLA-C02 is associate level for engineers who build and operate ML systems. AIF-C01 is foundational, for people who use AI on AWS without building it.

Neither is a prerequisite for the other.

Side by side

MLA-C02AIF-C01
LevelAssociateFoundational
Passing score720 / 1000700 / 1000
Duration170 min (beta)90 min
Price75 USD (beta)100 USD
Question typesMultiple choice, multiple responsePlus ordering and matching
Expected experience1+ year with SageMaker AI and Bedrock, plus 1+ year in a related roleUp to 6 months exposure
You are expected toBuild, deploy and operateUse and understand
Validity3 years3 years

During beta, MLA-C02 is actually cheaper than AIF-C01 despite being the harder exam. That is temporary.

The domains show the split

MLA-C02 — engineering:

DomainWeight
Data Preparation for ML and AI28%
ML Model and FM Development24%
Deployment and Orchestration24%
Operating, Monitoring, and Securing24%

AIF-C01 — understanding:

DomainWeight
Fundamentals of AI and ML20%
Fundamentals of GenAI24%
Applications of Foundation Models28%
Guidelines for Responsible AI14%
Security, Compliance, and Governance14%

AIF-C01 asks which service fits a business problem. MLA-C02 asks how you ingest and transform the data, select and tune the model, deploy the endpoint, configure auto scaling, wire the CI/CD pipeline and monitor it in production.

The out-of-scope lists settle it

AWS publishes what each candidate is not expected to do.

Out of scope for AIF-C01: coding models, data and feature engineering, hyperparameter tuning, building pipelines, statistical analysis, implementing security, developing governance frameworks.

Out of scope for MLA-C02: architecting full end-to-end solutions, setting ML strategy, integrating a wide array of new tools, deep expertise in multiple ML domains.

Nearly everything excluded from AIF-C01 is squarely inside MLA-C02. Meanwhile MLA-C02’s exclusions are about seniority and breadth, not capability — that territory belongs to the specialty and architect tiers.

Which should you take?

Take MLA-C02 if you write code against SageMaker AI or Bedrock, prepare training data, deploy endpoints, build pipelines, or operate models in production. AWS’s one-year expectation is realistic rather than decorative — this exam is difficult without hands-on experience.

Take AIF-C01 if you are a product manager, analyst, consultant, or a developer who consumes AI services. Also take it first if you are new to AI on AWS and want a structured foundation — at 100 USD it is a cheap ramp.

Take both, in order, if you are moving from using AI to building it.

If you can only take one, and you qualify for both

Take MLA-C02. It is associate level, it signals capability rather than literacy, and it covers the foundational material along the way. AIF-C01’s value is highest for people who will not go further; if you are going further, it is a stepping stone you can skip.

The exception is timing: during beta, MLA-C02 costs 75 USD. That makes sitting both unusually affordable, and there is a reasonable argument for taking AIF-C01 as a warm-up if your experience is closer to the one-year line than comfortably past it.