AI-103: Developing AI Apps and Agents on AzureDemo

By Microsoft · Question Mode

QuestionQ60

Implement generative AI and agentic solutions

You have a Microsoft Foundry project containing a deployed chat model.

A Python service sends API requests to the model. The service is integrated with an automated validation system that compares the generated output against approved response patterns.

Stakeholders report that minor wording differences are causing validation mismatches.

You need to update the request parameters to improve output stability while maximizing reasoning quality. How should you complete the Python code?

Select
message = client.messages.create(
    model="deployment-name",
    messages=[
        {"role": "user", "content": "Summarize the release notes in 3 bullet points."}
    ],
    max_tokens=800,
    temperature=,
    thinking={"type": "enabled"},
    output_config={"effort": }
)
Explanation

Setting temperature to 0 removes sampling randomness so the model's output becomes as deterministic as possible for the same input, which eliminates the small wording variations that were breaking pattern-based output validation. Setting the reasoning effort to "high" allocates the model the greatest amount of reasoning computation before it produces a response, which satisfies the requirement to maximize reasoning quality; "low" and "medium" trade reasoning depth for speed and would not maximize quality, and temperature values above 0 increase output variability rather than stability.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!