The AWS Certified AI Practitioner (AIF-C01) exam splits its 50 scored questions across five content domains: Applications of Foundation Models at 28%, Fundamentals of Generative AI at 24%, Fundamentals of AI and ML at 20%, and both Guidelines for Responsible AI and Security, Compliance, and Governance for AI Solutions at 14%.
Last updated August 2026.
The weights below describe how the 50 scored questions are distributed. The exam serves 65 questions in total, but 15 are unscored trial questions that are not identified while you sit it.
| AIF-C01 | |
|---|---|
| Level | Foundational |
| Total questions | 65 (50 scored, 15 unscored) |
| Time limit | 90 minutes |
| Passing score | 700 (on a 100 to 1,000 scale) |
| Scoring model | Compensatory (overall score, not per domain) |
| Cost | $100 USD |
All figures from the official AWS Certified AI Practitioner (AIF-C01) exam guide.
Percentages stay abstract until you convert them. Applying each weight to the 50 scored questions gives a far more concrete picture of where this exam actually spends its time.
| Domain | Weight | Approx. scored questions |
|---|---|---|
| 3. Applications of Foundation Models | 28% | About 14 |
| 2. Fundamentals of Generative AI | 24% | About 12 |
| 1. Fundamentals of AI and ML | 20% | About 10 |
| 4. Guidelines for Responsible AI | 14% | About 7 |
| 5. Security, Compliance, and Governance for AI Solutions | 14% | About 7 |
Domain weights from the official AIF-C01 exam guide. The question counts are those weights applied to 50 scored questions, so treat them as close approximations rather than a published blueprint.
The most useful thing the weights tell you: Domains 2 and 3 together are 52% of scored content, roughly 26 of the 50 questions. AIF-C01 is not a general AI exam with a generative AI chapter bolted on, it is a generative AI exam with an AI and ML foundation underneath.
That changes what preparation looks like. Foundation models, prompt engineering, the difference between fine-tuning and retrieval-augmented generation, and how to judge a model's output are not advanced extras, they are the bulk of the paper. Amazon Bedrock recurs across both domains as the service delivering these capabilities.
Responsible AI and Security, Compliance, and Governance are the lightest domains individually, but together they are 28% of scored content, about 14 questions. That is the same footprint as Applications of Foundation Models on its own.
They are also the cheapest marks available. The material is conceptual rather than technical (bias and fairness, transparency and explainability, data protection and access control for AI workloads), and the named AWS features attach cleanly to specific concerns: Guardrails for Amazon Bedrock to output safety, SageMaker Clarify to bias and explainability. A candidate fluent in generative AI who skipped these two domains is leaving roughly 14 questions on the table.
Because scoring is compensatory, you are aiming at a single scaled total of 700, not five separate bars. In practice that means no domain has to be perfect, but none can be blank either.
AIF-C01 has five content domains: Fundamentals of AI and ML (20%), Fundamentals of Generative AI (24%), Applications of Foundation Models (28%), Guidelines for Responsible AI (14%), and Security, Compliance, and Governance for AI Solutions (14%). The percentages are shares of scored content, not fixed question counts.
Applications of Foundation Models is the largest at 28% of scored content, roughly 14 of the 50 scored questions. It covers prompt engineering, fine-tuning versus retrieval-augmented generation, and evaluating how a foundation model performs.
Just over half. Fundamentals of Generative AI (24%) and Applications of Foundation Models (28%) together make up 52% of scored content, about 26 of the 50 scored questions. Generative AI is the core of this exam, not a side topic.
No. AWS uses a compensatory scoring model, so you are graded on one overall scaled score against the 700 passing mark rather than domain by domain. A strong domain can offset a weaker one, though a domain you skip entirely is hard to compensate for.
The guide lists the task statements and in-scope services behind every domain above. Questions are written directly from them, so it doubles as a blind-spot checklist.
Open the AWS exam guide →