AI is now the most common topic in testing job descriptions, and ISTQB offers two separate specialist certificates around it. They sound similar, but they answer different questions. One asks how you use generative AI to test better. The other asks how you test a product that contains AI.

Exam details in this article were checked against istqb.org in October 2026.

The one-line difference

  • CT-GenAI (Certified Tester Testing with Generative AI) teaches testers to use large language models (LLMs) and GenAI tools in their own work: writing prompts for test design, generating test data, reviewing outputs and managing the risks.
  • CT-AI (Certified Tester AI Testing) v2.0 teaches testers to test AI-based systems: machine-learning models, data quality, model behaviour and, from version 2.0, generative AI and LLM-based products.

ISTQB made the split sharper in 2026. Its CT-AI v2.0 page says content about using AI for testing was removed from that syllabus so it could focus only on testing AI-based systems. That material now belongs to CT-GenAI.

Side-by-side comparison

CT-GenAICT-AI v2.0
FocusUsing GenAI to support testingTesting AI-based systems
Current versionv1.1 (minor update announced 27 April 2026)v2.0 (announced 21 April 2026)
Questions4040
Points and pass mark46 points, 30 to pass (65%)65% pass mark
Duration60 minutes (+25% for non-native language)60 minutes (+25% for non-native language)
PrerequisiteISTQB CTFLISTQB CTFL
Chapters57, organised around the ML lifecycle
Typical audienceTesters, test analysts, automation engineers, test managers, developersTesters on ML, data or LLM products; QA for AI features

What CT-GenAI covers

ISTQB lists five chapters for CT-GenAI:

  1. Introduction to GenAI for software testing
  2. Prompt engineering for effective software testing
  3. Managing risks of GenAI in software testing
  4. LLM-powered solutions for software testing
  5. Deploying and integrating GenAI in test organisations

The business outcomes include understanding GenAI concepts, building prompting skills, identifying risks such as hallucinations and data leakage, and contributing to an organisation’s GenAI adoption roadmap. That last point matters in Indian IT services firms, where many testing teams are being asked to “add AI” to delivery without a clear plan.

What CT-AI v2.0 covers

ISTQB’s summary of the v2.0 changes:

  • The syllabus shrinks from 11 to 7 chapters, reorganised around the machine-learning lifecycle.
  • New coverage of generative AI and large language models as systems under test.
  • More emphasis on hands-on exercises and practical skills.
  • Alignment with emerging AI quality standards, including ISO/IEC 25059.

The exam pattern did not change: 40 questions, 60 minutes, 65% to pass. No prior machine-learning expertise is required beyond CTFL, although basic comfort with data and statistics helps a lot.

What happens to CT-AI v1.0?

ISTQB says v1.0 English exams and training continue until 21 April 2027, and other languages until 21 October 2027. Existing v1.0 certificates remain valid, and there is no bridging exam. If you are starting now, study v2.0.

Which one should you take?

Choose CT-GenAI if

  • You are a manual or automation tester who already uses tools like ChatGPT, Copilot or IDE agents to write test cases or scripts, and you want a structured method.
  • You lead a QA team that is piloting AI test generation and needs a shared vocabulary for risks and review.
  • You want the quicker win. It applies to almost every testing job.

Choose CT-AI v2.0 if

  • You test chatbots, recommendation engines, fraud models, OCR or any LLM-powered feature.
  • You work in a GCC or product company building AI features, where bias, robustness and data quality are part of acceptance.
  • You want to move towards AI assurance or model validation roles.

Taking both

The two do not overlap much now, so taking both is reasonable. A common order is CT-GenAI first, because it improves your day-to-day productivity, then CT-AI if your product work moves into AI features.

Where these fit with other learning

Neither certificate teaches a specific tool. Vendors change their AI features every few months, while the ISTQB syllabi focus on concepts that last longer: how to judge AI output, how to manage risk and how to evaluate a model. Use the certificate for the concepts. Learn the tools hands-on through official documentation, such as Playwright’s built-in test agents or the Selenium project’s guidance on using AI coding agents. Then you can talk about both in interviews.

How to prepare

  1. Download the official syllabus and sample exam from istqb.org. Both are free.
  2. Read every learning objective and note its K-level (remember, understand or apply). Apply-level objectives turn into scenario questions.
  3. For CT-GenAI, practise prompts on real artefacts: turn a user story into test conditions, generate boundary-value data, then check the output against the syllabus risk list.
  4. For CT-AI, revise basic ML terms (training, validation and test data, overfitting, bias, metrics such as precision and recall) before starting the syllabus.
  5. Take the sample exam under timed conditions. Then confirm the fee and exam provider with the Indian Testing Board before booking.

On cost, a third-party India guide (istqb.guru, May 2026) lists specialist exam fees of about ₹7,434 including GST for a first attempt through one remote provider. Confirm the current figure with ITB, because fees vary by provider and mode.

Frequently asked questions

Do I need coding or machine-learning experience for these exams?

No formal experience is required for either exam beyond holding CTFL. ISTQB says CT-AI v2.0 needs no prior AI or ML expertise beyond Foundation Level. Hands-on practice with AI tools will make CT-GenAI much easier to apply at work, though.

Will CT-GenAI help me get an AI testing job in India?

It shows a structured understanding of GenAI in testing, which hiring managers increasingly ask about. On its own it will not replace practical evidence. Pair it with a portfolio project, such as an AI-assisted test suite you reviewed and fixed, and explain your review process in interviews.

Is CT-AI v2.0 harder than v1.0?

ISTQB kept the same exam format, but the content is more practical and now includes generative AI systems. Candidates without exposure to ML projects should allow extra time for chapters on data and model evaluation.

Official sources

One response to “ISTQB CT-GenAI vs CT-AI v2.0 (2026): Which AI Testing Certification Should You Take?”

  1. […] 40-question, 60-minute exams according to ISTQB’s exam structure tables. Our comparison of CT-GenAI vs CT-AI explains which suits […]

Leave a Reply

Trending