Session 5· 12 questions

End-of-session quiz

What you'll learn

▸Confirm you understand output formatting, tool calling, and parse methods
▸Identify any gaps before moving to Session 6
▸Score 9+ out of 12 to move on confidently

12 questions across all three phases. Aim for 9 or more before moving to Session 6. If you miss one, go back to the lesson the answer explanation points to.

Phase 1 — Output formatting

Knowledge Check

What does response_format={"type": "json_object"} guarantee?

Code Check

Which flag enforces the schema during token generation?

Knowledge Check

When should you use response_format over tools?

Code Check

What does this Pydantic-free schema field mean: "type": ["string", "null"]?

Phase 2 — Function calling

Knowledge Check

What does tool_calls on the assistant message represent?

Code Check

After running a tool, what message role do you use to send the result back?

Code Check

Why must tool_call_id match the id on the tool_call?

Knowledge Check

What is the exit condition of a tool-call loop?

Knowledge Check

What should you do when a tool raises an exception?

Phase 3 — SDK parse methods

Code Check

What is the type of response.output_parsed when you use responses.parse(text_format=CalendarEvent)?

Code Check

Where does the typed object live when you use chat.completions.parse()?

Knowledge Check

Why is Pydantic + parse() cleaner than hand-written JSON Schema?

Scored 9+ out of 12?

You are ready for Session 6 — LangChain multi-provider. If you missed more than 3, revisit the matching lessons (each answer explanation points to the right one) before moving on.