- The Number: 700 on a 1-1000 Scale
- Why 700 Is Not a Raw 70%
- The Exam That Produces Your Score
- Where Your Points Come From: Six Domains
- What We Know and Don't Know About Weighting
- Make Sure You Are Studying the Current Outline
- A Domain-Ordered Readiness Plan
- Fees, Vouchers and What a Failed Attempt Costs
- After You Pass: Validity and Recertification
- Frequently Asked Questions
- The passing score is 700 on a 1-1000 scaled score, not a raw 70% requirement.
- Expect roughly 45-50 questions in a 50-minute timed window, per the exam-specific listing.
- Certiport does not publish pass rates, so any "official" pass-rate figure should be treated with suspicion.
- Study the revised six-domain outline, effective in English from June 30, 2026, not the legacy five-domain version.
The Number: 700 on a 1-1000 Scale
The passing score for the Information Technology Specialist - Artificial Intelligence certification is 700 on a scale of 1 to 1000. That single figure is published in Certiport's scoring policy, and it is the only passing threshold that matters. The exam is issued by Pearson and administered through Certiport, a Pearson VUE business, and the public credential name is IT Specialist - Artificial Intelligence (Pearson catalogue identifier INF-307).
If you are still orienting yourself on what this credential is, our overview pages on what ITS-AI certification is and what ITS-AI stands for cover the basics. This article focuses narrowly on the score: what 700 means, how it relates to the exam's format and content, and how to plan so you land comfortably above it.
Why 700 Is Not a Raw 70%
The most common misreading of "700 out of 1000" is to treat it as 70 percent correct. It is not. Certiport describes the result as a scaled score, which means your raw performance is converted onto the 1-1000 reporting scale. Scaling exists so that results remain comparable even when candidates see different sets of questions.
This matters for ITS-AI in particular because the exam is estimated at 45-50 questions, which is a variable range rather than a fixed form count. When the number of questions you see can differ, a simple "percent correct" mental model breaks down. Two candidates can have different raw tallies and still be measured against the same 700 standard.
What this means for your planning
- Do not reverse-engineer a number of questions you can miss. Certiport does not publish a conversion from raw answers to scaled score, so any "you can miss N questions" claim is speculation.
- Do not rely on practice-test percentages as a direct predictor. A 75% on a practice bank is encouraging, but it is not a guarantee of a 700 scaled result, because practice banks are not calibrated to the live exam's scaling.
- Treat every question as worth earning. Without a published conversion, the safest strategy is consistent accuracy across all six domains rather than gambling on skipping a weak area.
For a deeper look at how demanding the exam feels in practice, see How Hard Is the ITS-AI Exam?
The Exam That Produces Your Score
Your score comes from a computer-based, proctored knowledge assessment. Understanding the delivery mechanics helps you avoid losing points to logistics rather than knowledge.
| Element | What the sources state |
|---|---|
| Passing score | 700 on a 1-1000 scaled score |
| Question count | Estimated 45-50 questions (a variable range, not a fixed count) |
| Time limit | 50-minute maximum of timed exam time |
| Delivery | Pearson VUE test centers or OnVUE online proctoring; separate Certiport Authorized Testing Center (CATC) channel |
| Published pass rate | None; Certiport does not publish pass rates |
The 50-minute clock
The 50-minute limit refers to timed exam time. Certiport's exam-lengths page indicates that administrative steps and tutorials are not counted against it. With roughly 45-50 questions, that works out to a little over a minute per question on average, so pacing is a real consideration, especially for scenario-style items that require you to interpret a described AI situation before answering.
A note on older tutorial figures
The IT Specialist Exam Tutorial (copyright 2021) describes how to navigate and select answer options, and it mentions a generic 36-40 question estimate. That figure is older than the exam-specific listing and applies to the program generally, so for ITS-AI use the 45-50 question estimate and the 50-minute limit instead. Use the tutorial for interface behavior, not for sizing the exam.
Online versus test center
If you test through OnVUE, expect a system check and a live webcam proctor. Resolve technical checks well before exam day so that nothing interrupts your timed session. If you test through a CATC, expect photo and signature ID, an NDA, parental consent for candidates under 18, and a rule against personal paper or communication devices in the room.
Where Your Points Come From: Six Domains
The revised exam objectives define six domains. Your scaled score reflects performance across all of them, so a weak spot in one area can pull down a result that your strengths elsewhere would otherwise carry. The six headings are official exam objectives, not a preparation curriculum, so treat them as the scope of what can be tested.
Domain 1: AI Problem Definition
Frames whether and how AI should be applied to a problem in the first place.
- Assessing feasibility before committing to a build
- Translating a business need into a well-formed AI task
- Recognizing the human-interaction dimension of AI solutions
Domain 2: Data Collection and Transformation
Covers acquiring, preparing and shaping data into a usable state.
- Sourcing and assembling datasets
- Cleaning and transforming raw data
- Understanding how data quality affects downstream models
Domain 3: Feature Engineering
A separate domain in the current outline, covering how raw data becomes model inputs.
- Selecting and constructing meaningful features
- Handling the representation of different data types, including text for NLP work
Domain 4: AI Algorithms and Models
The core modeling content, including how to evaluate what you build.
- Choosing appropriate algorithms and model types
- Model-evaluation metrics and what they tell you
- Natural language processing and explainability concepts
Domain 5: Application Integrity and Deployment
Domain 5 uses this exact wording in the revised PDF, and application integration remains objective 5.1.
- Integrating models into applications
- CI/CD practices for AI systems
- Security considerations for deployed AI
Domain 6: Maintaining and Monitoring AI in Production
What happens after launch, which is easy to under-study.
- Detecting and responding to drift
- Ongoing monitoring of model behavior
- Decommissioning AI systems responsibly
For a fuller breakdown of each area, read our complete guide to all 6 ITS-AI content areas.
What We Know and Don't Know About Weighting
Candidates often ask which domain is "worth the most" toward the 700. The honest answer is that no official percentage weights are published for the six domains. The outline header simply counts the six domain lines. Some third-party practice products display item counts per heading, but those retail practice-bank counts are not official exam weights, and some of those products still use legacy headings.
This is exactly why the 700 threshold rewards breadth. With a 45-50 question exam spread across six domains, you will encounter every area, and you cannot assume any domain will be skipped or lightly sampled.
Make Sure You Are Studying the Current Outline
A subtle way to lose points before you ever sit down is to study outdated scope. The English version of the updated exam content took effect on June 30, 2026, as listed in Certiport's exam content updates archive, following an announcement dated April 29, 2026. Rollout dates for other languages differ, so the release statement here concerns English.
There is a documented source conflict worth knowing about. Pearson's general IT Specialist resources page still links a legacy copyright-2025 outline with five domains. That legacy outline should not be treated as complete current English coverage. The revised document is the one that separates out feature engineering, natural language processing, explainability, human interaction, feasibility, model-evaluation metrics, CI/CD, security, drift and decommissioning as distinct coverage areas.
Key Takeaway
Before you buy any study material, check whether it is built around six domains including Feature Engineering and Maintaining and Monitoring AI in Production. Material organized around five domains likely reflects the legacy outline and may leave gaps against the current exam.
Our ITS-AI study guide is organized around the current scope, and you can pressure-test your coverage with timed questions on the main ITS-AI practice test site.
A Domain-Ordered Readiness Plan
The program suggests approximately 150 hours of preparation, and the objective profiles describe at least 150 hours. These are readiness profiles, not a verified logged-hours application requirement, so nobody checks your hours, but they are a useful benchmark for how much ground the six domains cover. There is no degree requirement beyond knowledge of the objective domains; see ITS-AI requirements for the full eligibility picture.
Ordering your study to follow the lifecycle of an AI project makes the content stick, because each domain feeds the next.
Problem Definition and Data
- Work through Domain 1 feasibility and framing scenarios
- Cover Domain 2 data collection and transformation, since every later domain depends on it
Features and Models
- Study Domain 3 feature engineering as its own topic, not a footnote to modeling
- Begin Domain 4: algorithms, model-evaluation metrics, NLP and explainability
Deployment and Production
- Cover Domain 5 integration, CI/CD and security
- Cover Domain 6 drift, monitoring and decommissioning, which candidates tend to underestimate
Timed Practice Across All Six
- Take timed sets sized to the 45-50 question, 50-minute format
- Revisit whichever domain your results show as weakest
If you want a condensed review once you have covered the material, our ITS-AI cheat sheet distills the must-know facts.
Fees, Vouchers and What a Failed Attempt Costs
Because the score is a hard gate, it helps to understand the money attached to each attempt. The two voucher routes differ in important ways.
| Option | Price | Key conditions |
|---|---|---|
| Exam-only voucher (Pearson MindhubPro) | USD 127 web/member pricing | Country-restricted; unavailable in listed markets; expires 12 months after purchase |
| CATC voucher with retake | USD 133 | US-only; proctoring may cost extra; not usable at Pearson VUE centers or through OnVUE |
Two points deserve emphasis. First, the voucher's 12-month expiry is separate from the credential's validity period, so buying early does not extend anything about the certification itself. Second, the CATC voucher bundles a retake, which can matter if you want a safety net, but it only works through the Certiport Authorized Testing Center channel, not at Pearson VUE centers or via OnVUE.
For the full financial picture, including how these prices fit into total spend, see our ITS-AI certification cost breakdown. If you are still deciding whether the investment makes sense, our analysis of whether the ITS-AI certification is worth it walks through the trade-offs.
After You Pass: Validity and Recertification
Reaching 700 earns a credential with a defined lifespan. For awards on or after June 18, 2025, validity is five years. Earlier awards remain nonexpiring. When a five-year credential lapses, recertification requires a new voucher and passing the current exam, which means you would be measured against whatever the then-current objectives are, not the outline you originally studied.
That is one more reason to keep your knowledge of the production-side topics sharp, since the field's emphasis on deployment, security and monitoring is exactly where the revised outline expanded. Candidates thinking about where the credential can lead can explore ITS-AI jobs and the ITS-AI salary guide, and those scheduling their attempt should check ITS-AI exam dates and scheduling.
Frequently Asked Questions
The passing score is 700 on a scale of 1 to 1000. It is a scaled score, so it is not equivalent to answering 70% of questions correctly.
The exam-specific listing estimates 45-50 questions with a 50-minute maximum of timed exam time. The question count is a variable range, not a fixed form count, and administration and tutorial time are not counted in the 50 minutes.
Certiport does not publish pass rates, so no verified figure exists. Our ITS-AI pass rate article explains what can and cannot be said from available information.
No official percentage weights are published for the domains, so you cannot assume equal or unequal weighting. Plan for balanced coverage of all six, and spend extra time on your weakest areas.
No. The voucher expires 12 months after purchase, which is separate from credential validity. A credential awarded on or after June 18, 2025 is valid for five years.
Whatever your starting point, the 700 threshold is best approached by mastering all six domains of the current outline rather than chasing a number. Measure your readiness with realistic timed questions on the ITS-AI practice test site before you commit to an exam date.