Claude Certified Architect – Foundations Practice Question 57: Stop Sequences Delimiter
Claude Certified Architect, Foundations · Test 6 · question 7 of 10
A template asks Claude to write a product description and then emit the marker ---END---. The application only needs the text before the marker and does not want to pay for any tokens generated after it. The team also wants to tell programmatically that generation ended because the marker appeared, not because of a length limit. What should the request do?
Show answer and explanation
Answer:
B is correct. stop_sequences is a request parameter holding custom strings; when Claude emits one, generation halts there, the response's stop_reason is "stop_sequence", and the stop_sequence field says which one fired. Generation stops at the marker, so nothing after it is produced or billed.
A is wrong because a small max_tokens cuts output at an arbitrary point and yields stop_reason "max_tokens", and "end_turn" means Claude finished naturally. C is wrong because stop_reason is a response field you read, not a request parameter. D is wrong because it works for parsing but lets the model keep generating (and billing) past the marker, and "end_turn" cannot tell you the marker was reached.
Reference: Handling stop reasons