A test team is evaluating a tool that can take a screenshot of a bug and automatically generate a bug report with a title, steps to reproduce, and expected vs. actual results. The tool must understand the visual context of the screenshot and the textual context of the session log. Which model capability is essential for this tool?

Multimodal understanding to process both the screenshot and the log.

Text summarization to condense the log.

Text-to-image generation to recreate the bug.

Image captioning to describe the screenshot.

In software testing, which task is a potential application of a multimodal LLM?

Calculating test coverage metrics from code.

Analyzing UI screenshots to detect visual defects.

Compiling test code into machine code.

Managing test environment virtual machines.

A QA team is testing a new e-commerce site. They have a set of screenshots showing the product page at various stages of a purchase flow. They want to automatically generate test data (e.g., product names, prices, descriptions) from these screenshots to use in their test scripts. Which approach is the most effective?

Use an optical character recognition (OCR) tool to extract all text and then manually map it to the test script fields.

Use a multimodal LLM to extract the relevant text and structured data from each screenshot.

Use a screen-scraping tool to capture the HTML behind the screenshots.

Use a text-only LLM to generate random product data based on a prompt.

A QA team is testing a responsive web design. They have screenshots of the same page on mobile, tablet, and desktop. They want to use a multimodal LLM to verify that all key UI elements (e.g., navigation menu, search bar) are present and visible in all three versions. What is the most effective way to structure this task?

Use a pixel-diff tool to compare the screenshots and report any differences.

Use a text-only LLM to analyze the CSS media queries in the code.

Provide all three screenshots to the model and ask it to compare the presence of the specified elements across them.

Provide one screenshot at a time and ask the model to describe it, then manually compare the descriptions.

A test engineer is evaluating tools to automatically generate accessibility test cases from a web page's rendered output. The tool needs to understand both the visual layout and the underlying text content to suggest improvements. Which type of model is most appropriate for this task?

A speech-to-text model that processes the audio of a screen reader.

A vision-language model that can analyze the page screenshot and answer questions about element placement and text.

A text-only large language model that parses the HTML source code.

A traditional rule-based accessibility checker that scans the DOM.

Certified Tester Testing with Generative AI

Domain 1Objective 3

Multimodal LLMs and Vision-Language Models CT-GENAI Practice Questions (Page 2)

Part of the Introduction to Generative AI for Software Testing domain, which makes up ~12% of our current practice bank. ISTQB does not publish an official question count, but from its 60-minute exam (~25–40 total, ~3–5 in this domain), expect 1–1 from this objective — we provide 22 practice questions to prepare you well beyond it. (estimate)

22questions here

5free pages

5concepts

Questions 6–10

6application · medium
A test team is evaluating a tool that can take a screenshot of a bug and automatically generate a bug report with a title, steps to reproduce, and expected vs. actual results. The tool must understand the visual context of the screenshot and the textual context of the session log. Which model capability is essential for this tool?
Select an answer first
7foundation · easy
In software testing, which task is a potential application of a multimodal LLM?
Select an answer first
8application · medium
A QA team is testing a new e-commerce site. They have a set of screenshots showing the product page at various stages of a purchase flow. They want to automatically generate test data (e.g., product names, prices, descriptions) from these screenshots to use in their test scripts. Which approach is the most effective?
Select an answer first
9application · medium
A QA team is testing a responsive web design. They have screenshots of the same page on mobile, tablet, and desktop. They want to use a multimodal LLM to verify that all key UI elements (e.g., navigation menu, search bar) are present and visible in all three versions. What is the most effective way to structure this task?
Select an answer first
10application · medium
A test engineer is evaluating tools to automatically generate accessibility test cases from a web page's rendered output. The tool needs to understand both the visual layout and the underlying text content to suggest improvements. Which type of model is most appropriate for this task?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to CT-GENAI

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by ISTQB. “CT-GENAI” is a trademark of its owner, used for identification only.

Multimodal LLMs and Vision-Language Models CT-GENAI Practice Questions (Page 2)

Questions 6–10

In software testing, which task is a potential application of a multimodal LLM?

A test engineer is evaluating tools to automatically generate accessibility test cases from a web page's rendered output. The tool needs to understand both the visual layout and the underlying text content to suggest improvements. Which type of model is most appropriate for this task?

Keep studying — Introduction to Generative AI for Software Testing