- What Generative AI Foundations (GAF) Actually Is
- Who Issues It and How It Fits Into Certiport
- Exam Format, Timing, and Scoring
- The Four Exam Domains
- Who the Exam Is Built For
- Registration Paths and Voucher Costs
- Testing Rules and Validity Period
- What the Credential Signals to Employers
- A Domain-Ordered Preparation Plan
- Frequently Asked Questions
- GAF is Pearson's Generative AI Foundations exam, delivered through Certiport's Critical Career Skills (CCS) program.
- The exam runs 50 minutes and requires a scaled score of 700 on a 1-1000 scale.
- Four domains: methods, basic prompting, prompt refinement, and ethics, law, and societal impact.
- Certiport vouchers cost USD 72 (or USD 83 with a retake) and work only at Certiport Authorized Testing Centers.
What Generative AI Foundations (GAF) Actually Is
The acronym "GAF" is shared by several unrelated credentials, so it is worth being precise about the subject of this article. Here, GAF means the Generative AI Foundations certification: a proctored, computer-based knowledge assessment that verifies a candidate understands how generative AI tools work, how to prompt them effectively, how to improve their output, and how to use them responsibly. If you have landed here searching for a broader definition, our pages on what GAF is and what GAF stands for cover the naming question in more detail.
The exam is deliberately brand-agnostic. It does not test whether you can operate one vendor's chatbot or image generator. Instead, it checks conceptual understanding and practical judgment that transfer across tools: what a generative model does, how to phrase a request so the output is usable, how to iterate when the first answer misses, and where legal and ethical lines sit. That framing makes it an entry-level, foundations-tier credential rather than a developer or data-science certification.
Who Issues It and How It Fits Into Certiport
The certification is issued by Pearson and developed and administered through Certiport, a Pearson VUE business. Within Certiport's catalog it sits in the Critical Career Skills (CCS) family, a group of credentials aimed at workplace-readiness skills rather than deep technical specializations. Pearson announced the certification on October 1, 2024, and the objectives document that defines the exam content carries a 2024 copyright.
That lineage matters for two reasons. First, the exam follows Certiport's standard delivery conventions: scaled scoring, a built-in tutorial, review-before-submit navigation, and strict identity and proctoring rules. Second, the credential's value rests on the Pearson and Certiport brand names, which many schools, workforce programs, and employers already recognize from other certifications. If you want the broader picture of the credential itself, see our overview at GAF certification.
Exam Format, Timing, and Scoring
GAF is a proctored, computer-based knowledge assessment. Candidates select answers, can flag items they want to revisit, and can review their responses before final submission. The key numbers are straightforward:
| Element | Detail |
|---|---|
| Timed exam duration | 50 minutes (tutorial and administrative seat time are not counted) |
| Passing score | 700 on Certiport's 1-1000 scaled-score system |
| Question count | Varies by exam form; the Certiport tutorial specifies 40-45, while the Mindhub voucher page estimates 45-50 |
| Interaction style | Answer selection with flagging and review before submission |
| Delivery | Proctored, computer-based |
Why the question count is a range
Candidates often ask for one definitive question count, and the honest answer is that the issuer's own materials disagree slightly. Certiport's GAF tutorial describes 40-45 questions, while the Mindhub voucher page estimates 45-50. Certiport's scoring policy explains why: totals differ by exam form. Plan for a range rather than a fixed number, and do not build your pacing strategy around a single figure.
What "700 out of 1000" really means
The passing score is a scaled score, not a percentage of questions answered correctly. A 700 does not mean you must get 70 percent of items right; scaled scoring adjusts for differences between exam forms. For a deeper look at how to interpret this, see our dedicated page on the GAF passing score.
The Four Exam Domains
Certiport's objectives document organizes the exam into four domains. The outline lists these headings without published percentage weights, so you should not assume any domain counts more than another. Treat all four as testable, and use the complete guide to the four GAF content areas for a deeper breakdown of each.
Domain 1: Generative AI Methods and Methodologies
The conceptual foundation. Candidates need to understand what generative AI is, how it differs from other kinds of software, and the general approaches behind generating text, images, and other content.
- Core vocabulary: models, training data, outputs, and what "generative" implies
- How generative systems produce content versus how traditional rule-based tools behave
- Practical awareness of strengths, limits, and common failure modes such as inaccurate or fabricated output
Domain 2: Basic Prompt Engineering
The skill of asking well. This domain covers how the wording of a request shapes what a generative tool returns.
- Writing clear, specific instructions rather than vague requests
- Supplying context, role, format, and constraints so output matches the need
- Recognizing why two differently worded prompts produce very different results
Domain 3: Prompt Refinement
The skill of iterating. A first answer is rarely the best one, and this domain focuses on diagnosing a weak response and improving it.
- Identifying what is missing or wrong in an initial output
- Adjusting wording, adding detail, or narrowing scope in follow-up prompts
- Treating prompting as a loop of test, evaluate, and revise
Domain 4: Ethics, Law, and Societal Impact
The responsibility layer. Candidates must reason about the consequences of using generative AI at work and in society.
- Bias, fairness, and the risk of reproducing harmful patterns
- Copyright, attribution, privacy, and data-handling concerns
- The broader effects of AI on workplaces, trust, and information quality
Key Takeaway
Domains 2 and 3 reward practice rather than memorization. Spend real time writing prompts, judging the results, and rewriting them, because scenario-style questions are easier when you have done the iteration yourself.
Who the Exam Is Built For
The issuer describes the exam as best suited for candidates ages 14 and older, which signals a very low barrier to entry. There is no bachelor's degree requirement. What Certiport does recommend is preparation of roughly 150 combined hours of instruction and hands-on use of generative AI tools, along with familiarity with productivity applications. Those hours describe suggested candidate preparation; they are not the exam timer and not a mandatory course length. Our GAF requirements guide walks through eligibility in more depth.
In practice, the audience tends to fall into a few groups:
- Students in secondary school, college, or workforce-readiness programs who want a recognized credential showing AI literacy.
- Career changers and job seekers who want documented evidence of generative AI skills on a resume.
- Working professionals in administrative, marketing, operations, or office roles who already use productivity software and are being asked to use AI tools responsibly.
- Educators and training programs looking for a standardized benchmark to certify learners.
If you are weighing difficulty before committing, our analysis of how hard the GAF exam is helps set expectations.
Registration Paths and Voucher Costs
Pricing and booking mechanics are where candidates most often get tripped up, because there is more than one voucher type and they are not interchangeable. The facts below come from the issuer's store and voucher pages.
| Option | Price (U.S.) | Where it works |
|---|---|---|
| Certiport single-exam voucher | USD 72 | Certiport Authorized Testing Centers (CATC), in person or remote |
| Certiport voucher with retake | USD 83 | Certiport Authorized Testing Centers (CATC), in person or remote |
| Pearson Mindhub voucher | USD 85 | Subject to country restrictions; must match the country selected at checkout |
Watch for extra charges and channel limits
Certiport vouchers are valid only for CATC delivery and cannot be redeemed through OnVUE or a Pearson VUE Testing Center. A CATC may also add its own proctoring charge on top of the voucher price. The Mindhub voucher is listed at USD 85, and Mindhub Pro lists both web and member prices as USD 85. Mindhub vouchers carry country restrictions, so the country you select at checkout must match your eligibility.
For a fuller view of what you may end up paying, see our GAF certification cost breakdown, and for scheduling logistics, the GAF exam dates and scheduling guide.
Testing Rules and Validity Period
Proctoring and identity rules
Certiport Authorized Testing Centers require proctoring and valid identification. Candidates under 18 need parental consent. Books and personal devices must remain inaccessible during the exam, so the experience is closer to a controlled assessment than an open-resource task, even though the subject itself is AI.
How long the credential lasts
CCS certifications earned on or after June 18, 2025 are valid for five years. Recertification requires purchasing a new voucher and passing the current exam. Certifications earned earlier retain their existing, nonexpiring status. Because generative AI changes quickly, the five-year window is a sensible acknowledgment that the skills and the tools will evolve.
What the Credential Signals to Employers
Generative AI Foundations is an entry-level, vendor-neutral credential, so it is best understood as evidence of baseline AI literacy rather than proof of advanced technical ability. It does not make someone a machine-learning engineer. What it communicates is that the holder understands the basic concepts, can prompt and refine outputs with intent, and has been exposed to the legal and ethical considerations of using these tools at work.
That signal is most relevant in roles where generative AI is a productivity tool rather than the product itself: administrative and operations positions, marketing and communications, customer support, education and training, and general office roles where staff are increasingly expected to use AI assistants responsibly. Because the exam includes an explicit ethics, law, and societal impact domain, it also speaks to employers who worry about careless use of AI, such as pasting sensitive data into public tools or publishing unchecked output.
We deliberately avoid quoting salary figures here, since no reliable, GAF-specific earnings data is published by the issuer. If you want a qualitative discussion of how the credential may affect pay and opportunity, our GAF salary guide, our look at GAF-relevant jobs, and the ROI analysis on whether GAF is worth it take that up carefully.
A Domain-Ordered Preparation Plan
Rather than generic study advice, the most useful way to organize preparation is to follow the logical dependency between the four domains. Each builds on the one before it. A four-week arrangement works well for someone who already uses productivity software and has some hands-on time with AI tools; stretch it out if you are newer to the subject. For a longer treatment, see the GAF study guide.
Domain 1: Methods and Methodologies
- Learn the core vocabulary so later domains make sense
- Note how generative output can be wrong, biased, or invented
Domain 2: Basic Prompt Engineering
- Write prompts that specify context, role, format, and constraints
- Compare vague versus specific wording on the same task
Domain 3: Prompt Refinement
- Diagnose weak outputs and revise prompts in several rounds
- Practice explaining why a revision improved the result
Domain 4: Ethics, Law, and Societal Impact, plus review
- Work through bias, copyright, privacy, and data-handling scenarios
- Take timed practice questions covering all four domains
Because the real exam is timed at 50 minutes, finish with timed practice so that pacing feels natural. Our GAF practice tests let you rehearse under similar conditions, and the GAF cheat sheet is useful for a last-day review of must-know facts. For structured learning options, see our page on GAF training. If you also want to gauge your readiness against typical outcomes, our discussion of the GAF pass rate explains what is and is not publicly known.
Frequently Asked Questions
On this site, GAF refers to Generative AI Foundations, a Pearson certification developed and administered through Certiport within its Critical Career Skills program. It is unrelated to other credentials that happen to share the same acronym.
The timed portion is 50 minutes, excluding the tutorial and administrative seat time. The passing score is 700 on Certiport's 1-1000 scaled-score system, which is not the same as a percentage of questions answered correctly.
It depends on the exam form. Certiport's tutorial specifies 40-45 questions, while the Mindhub voucher page estimates 45-50. Certiport's scoring policy states that totals differ by form, so plan for a range.
No bachelor's degree is required, and the exam is described as best suited for ages 14 and older. The issuer expects about 150 combined hours of instruction and hands-on use of generative AI tools, plus familiarity with productivity applications, as preparation rather than a formal prerequisite.
CCS certifications earned on or after June 18, 2025 are valid for five years. Recertifying requires a new voucher and passing the current exam, while earlier nonexpiring certifications keep their existing status.
Generative AI Foundations is a compact, accessible credential: a 50-minute proctored exam, four clearly named domains, and a modest voucher cost. If the subject is new to you, start with the concepts in Domain 1, build real prompting habits in Domains 2 and 3, and treat Domain 4 as seriously as the technical material. When you are ready to test yourself, head to the main practice test site.