The ISTQB CT-AI v2.0 exam has 40 questions, 60 minutes, and 29 of 44 points (65%) to pass. ExamCaliber has 220 free ISTQB CT-AI v2.0 practice questions in 5 full-length mock exams of 40 questions, 60 minutes and 65% to pass — every answer, right and wrong, carries a written rationale. No sign-up, no paywall.
40 questions, worth 44 points in total. The questions are not equally weighted: some are worth one point and some two, with the heavier ones at the higher cognitive levels where you apply or analyse rather than recall.
29 of 44 points, which is 65%. Because the weights differ, track points rather than the number of questions: missing four two-point items costs as much as missing eight one-point items.
60 minutes, or 75 minutes if you qualify for the 25% non-native-language allowance. Request it when you book — it is granted at booking, not at the exam desk.
Yes. The ISTQB Certified Tester Foundation Level certificate is a prerequisite for CT-AI v2.0. This is the specialist exam where candidates most often discover the prerequisite late, so check it before you plan a date.
ISTQB does not publish a per-chapter question count for CT-AI. For CTFL there is an official exam structure table; for CT-AI there is not. Any provider telling you "chapter 3 is worth eight questions" has invented the figure. What you can plan against is the shape of the syllabus rather than a fixed split.
Seven areas, organised around the lifecycle of an AI-based system: introduction to AI; quality characteristics for AI-based systems and their acceptance criteria; machine learning (including data, functional performance metrics for classification, and neural networks); testing AI-based systems including generative AI and large language models; input data testing; model testing; and machine learning development testing.
They are different certifications and people book the wrong one. CT-AI is about testing AI-based systems — you are the tester and the AI is the system under test. CT-GenAI is about using generative AI to support your testing work — the AI is your tool. If your job is to assure a model someone else trained, you want CT-AI.
It is being retired. The English v1.0 exam remains available until 21 April 2027 and non-English versions until 21 October 2027, after which all v1.0 exams and training courses end. v2.0 reorganised the material substantially, so study v2.0 material only — a v1.0 question bank tests topics that have moved or gone.
No. There is no programming test and no data science qualification required. You do need to read a confusion matrix, compute and interpret classification metrics, and understand how an ML workflow produces a model — conceptual fluency, not implementation skill.
Yes, expect it. Given a confusion matrix you should be able to compute precision, recall and F1 and say which one matters for the stated goal. Know the accuracy paradox: 99% accuracy on a rare-class problem can mean the model never predicts the rare class at all.
For a tester who already holds CTFL, three weeks at six to eight hours a week is a realistic plan, with the middle week spent on machine learning and metrics. If ML is entirely new to you, allow four to five weeks — the vocabulary and the metric arithmetic take the extra time, not the testing concepts.
No. ISTQB certifications are awarded for life and need no renewal, recertification or continuing-education credits. A certificate earned against v1.0 stays valid even after v1.0 exams are withdrawn.
ISTQB does not set a single global price. The fee is set by the national member board or the accredited exam provider you book with, so it varies by country and by whether you sit remotely or at a test centre. Check your local member board's site for the current figure.
No, and nobody legitimately has those — ISTQB does not release live exam questions. Ours are original questions written against the published v2.0 syllabus, with a written rationale on every answer option including the wrong ones. One difference worth knowing: our mocks weight every question at one point, while the real exam mixes one- and two-point items, so read your result as a percentage rather than mapping it onto the 29-of-44 threshold.
80% or better across three different mocks, taken timed, with no topic area consistently below 65%. A single pass at the threshold is not a readiness signal — it is the pass mark with nothing left over for a bad morning.