Free CT-GenAI mock exams

CT-GenAI is marked against the syllabus, not against how you use AI tools day to day. Hallucination, reasoning error, bias, few-shot prompting, meta prompting, RAG, fine-tuning and LLMOps all have precise syllabus meanings, and the distractors are usually ideas that sound right to anyone who has used a chatbot. When you miss a question, check the syllabus wording rather than your own habits.

Prompt engineering is the heaviest chapter, and it holds almost all of the two-point apply-level questions: given a test task, pick the prompt structure or technique that fits it, or spot what a prompt is missing. These questions reward practising with prompts, not reading about them, so rehearse them until you can name the missing part of a prompt quickly.

The risk chapter is the second biggest and the easiest to underrate. Expect output to judge: is this a hallucination, a reasoning error or a bias, and which mitigation fits? Privacy and security questions turn on what you send to which model, so keep the data-exposure risks and their mitigations distinct.

Infrastructure and adoption are small chapters with a lot of vocabulary: RAG versus fine-tuning, agents, LLMOps, shadow AI, choosing between an LLM and an SLM, and the adoption phases. They are easy marks if you can tell the terms apart and easy losses if you skimmed them.

The real exam gives two points for each apply-level question and one for the rest; our mocks count every question the same, so treat a mock score near the pass line as borderline and look at your prompt-engineering accuracy first.