Unverified50% confidenceFactExact time
If finish_reason is length while the client still treats the content as a complete answer, this falls under the truncation_ambiguity failure classification.
1
Sources
50%
Confidence
Long-term
Relevance
10/19/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedCommon values of the finish_reason field include stop (natural end) and length (truncated upon reaching the maximum token count).70% similarUnverifiedIn practice, stop_reason === 'tool_use' does not reliably reflect message structure: a response may contain a real tool call block yet report end_turn.63% similarUnverifiedOnly responses with a finish_reason of stop or tool_calls should be cached; responses truncated by length or those that errored should not be cached.61% similarUnverifiedPlaceholder content in the prompt goes directly into the agent's context window; if results_json is large, it may crowd out reasoning space or increase token consumption, so for batch failure scenarios it is recommended to truncate or summarize first.61% similarUnverifiedThe OpenAI Responses API triggers a response.incomplete event when a response ends early (such as hitting the max_output_tokens limit), and that event also carries response.usage; listening only for response.completed will miss the usage of truncated answers.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1066811API
curl https://kongchang.com/api/v1/knowledge/claims/1066811MCP
get_claim(id=1066811)