AWS AI Certifications Explained

AWS Certified Generative AI Developer – Professional (AIP-C01): the new top rung

The newest addition sits at Professional level. The AWS Certified Generative AI Developer – Professional opened for beta registration in November 2025 at a $150 beta price, with the standard version open for registration as of March 2026 at the usual Professional-tier $300 USD. It validates the ability to integrate foundation models into production applications, including work with Amazon Bedrock and Bedrock AgentCore, and AWS recommends around two years of cloud experience plus a year of hands-on generative AI work. AWS launched an "Agentic AI Demonstrated" microcredential alongside it.

Because AIP-C01 is so new, fewer format details are firmly established in public documentation than for the older exams — confirm question count, duration and passing score in the official AIP-C01 exam guide on aws.amazon.com/certification before booking. What is clear is its position: this is the first Professional-level AWS certification dedicated to AI, and it gives the AI path a genuine three-level ladder for the first time.

What happened to the Machine Learning Specialty (MLS-C01)?

For years, MLS-C01 was the AWS machine learning certification, sitting in the Specialty tier at $300 with a 180-minute exam. It is now retired: the last day to take it was 31 March 2026, and it can no longer be booked or renewed. If you already hold it, your certification remains active for three years from the date you earned it — it does not vanish — but when it lapses you would recertify into the new lineup instead, with MLA-C01/C02 and AIP-C01 as the effective successors AWS named in its portfolio announcement.

Study materials, archived MLS-C01 exam-topic resources and old blog advice referencing the Specialty are still everywhere online, which is exactly why it is worth being precise: any 2026 study plan that ends at MLS-C01 is pointing at an exam you cannot sit. The fuller story of what the Specialty covered and where its holders go next lives in the dedicated AWS Machine Learning Specialty guide.

AIF-C01 vs MLA-C01: which one is actually yours?

The most common decision inside this lineup is between the two 2024 launches. They share launch timing and question types but almost nothing else.

Choose AI Practitioner (AIF-C01) if your work involves understanding, evaluating, procuring or managing AI solutions rather than building them. It is shorter (90 minutes), cheaper ($100), conceptual, and forgiving of a non-engineering background. It is also a sensible first-ever AWS exam for someone whose interest in the platform starts with AI.

Choose ML Engineer Associate (MLA-C01) if you can already take a model from data preparation through training to a deployed, monitored endpoint — or you are close to that and willing to close the gap with hands-on SageMaker work. The exam is longer (130 minutes), costs $150, and expects practitioner depth in line with AWS's one-year experience recommendation.

Doing both, in order, is legitimate. Because every active AWS certification earns you a 50% discount voucher for your next exam, a Foundational-then-Associate sequence reduces the second exam's effective cost — and passing the Associate later renews a Foundational certification anyway under AWS's recertification rules. But nothing forces the sequence: there are no prerequisites, and an experienced ML engineer can go straight to MLA.

A three-question self-test that settles it for most people:

1. Have you deployed a model to production on AWS, or could you with documentation open? Yes → MLA. No → AIF first.
2. Does your role require you to build, or to decide and communicate? Build → MLA. Decide/communicate → AIF is likely sufficient on its own.
3. Is generative AI application development your day job? If yes, treat MLA as a waypoint and read the AIP-C01 exam guide now, because that Professional cert is the one that matches your work.

How the AI path connects to the rest of AWS's portfolio

The AI certifications live inside the same system as every other AWS credential, and four shared rules shape any plan you build around them.

Scoring is scaled, not percentage-based. Foundational exams pass at 700, Associate at 720, and Professional and Specialty at 750, on a 100–1000 scale. A 720 does not mean 72% of questions correct; AWS scales raw marks across exam forms, so no published number tells you "how many you can get wrong".

Everything expires in three years. Renewal is by passing the current version of the exam or a qualifying higher-level exam. AWS also lists a one-year maintenance extension for Associate and Professional certifications via an active AWS Skill Builder subscription — check the recertification page for current terms.

No exam has prerequisites. Experience levels are recommendations. You could, in principle, sit AIP-C01 with no other AWS certification — whether that is wise is a separate question from whether it is allowed.

The 50% voucher compounds. Each active certification halves the cost of the next exam, which is why multi-cert AI paths cost less in practice than list prices suggest.

Many AI-path candidates also hold or consider the general cloud certifications — Cloud Practitioner as a foundation or Solutions Architect Associate for breadth. Those are covered in their own guides on this site and are one sentence outside this article's scope: if you want the entry-level general exam explained from scratch, see the AWS Cloud Practitioner exam guide.

Preparing for an AWS AI exam: what changes at each level

A full study plan belongs to each exam's own guide, but the level structure itself tells you how preparation differs, and it is worth internalising before you commit.

At Foundational level (AIF-C01), preparation is mostly reading, watching and concept-checking: services, terminology, responsible-AI principles, and where each AWS AI service fits. At Associate level (MLA-C01), the exam guide expects things you can only really learn by doing — so hands-on time in SageMaker matters more than any amount of reading. At Professional level (AIP-C01), scenario judgement dominates: integrating foundation models into real applications, with the trade-offs that implies.

At every level, the same feedback loop applies once you have covered the material: test yourself under exam conditions, find the weak areas, and go back to them. Working through practice questions domain by domain — and reviewing why wrong answers are wrong rather than memorising which letter was right — is the reliable way to check understanding against the exam guide's objectives. ExamPractice's Amazon exam hub collects AWS exam-topic pages with free sample questions, and subscribers can run timed practice-test simulations to benchmark readiness before spending the booking fee.

Frequently asked questions

Is the AWS Machine Learning Specialty still available in 2026?

No. MLS-C01 retired on 31 March 2026 and can no longer be taken or renewed. Existing holders keep an active certification for three years from the date they earned it. The current alternatives are AI Practitioner (Foundational), Machine Learning Engineer – Associate, and Generative AI Developer – Professional.

Do I need AI Practitioner before Machine Learning Engineer – Associate?

No. AWS exams have no prerequisites, so you can sit MLA-C01 directly. AIF-C01 first makes sense if you are new to AWS's AI services or want a lower-stakes introduction to the exam experience — and holding it earns a 50% voucher toward the Associate exam.

Should I book MLA-C01 now or wait for MLA-C02?

If you can be ready before 28 September 2026 (the last day for MLA-C01 in English), booking the current version lets you study against a stable, well-documented exam guide. If not, wait for the official MLA-C02 guide rather than preparing against unconfirmed third-party claims about the update.

Are there hands-on labs in any AWS AI exam?

No. Current AWS exams are question-based, with no lab component. AIF-C01 and MLA-C01 do include the newer ordering, matching and case-study question types alongside multiple choice and multiple response.

How long is an AWS AI certification valid?

Three years, like all AWS certifications. You renew by passing the current exam version or a qualifying higher-level exam, and AWS lists additional maintenance options on its recertification page.

Choosing your rung on the ladder

AWS's AI lineup in 2026 is cleaner than it has ever been: one conceptual credential for people who work around AI (AIF-C01), one hands-on credential for people who build it (MLA-C01, becoming MLA-C02), and one Professional credential for people shipping generative AI applications (AIP-C01) — with the old Machine Learning Specialty formally retired. Match the level to your actual work rather than to the most impressive-sounding title, read the official exam guide for whichever rung you pick, and remember the deadline that matters this year: English-language MLA-C01 testing ends 28 September 2026. From there, it is ordinary exam preparation — cover the guide, drill your weak domains, and book when your practice results say you are ready.