Unverified50% confidenceFactExact time
The 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.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/17/2026
First Seen
Valid until: 1/15/2027
Sources
Related Entities
Related Claims
UnverifiedOpenAI 流式响应以 data: [DONE] 标记结束，Anthropic 则使用 event: message_stop 事件类型77% similarUnverifiedOpenAI identifies refusals and max-token truncation as two scenarios in which a response may not match the schema.74% similarUnverifiedPassing json.loads parsing alone cannot determine whether a model response was truncated, because a truncated response can still parse successfully if the brackets happen to close.70% similarUnverifiedUsing exponential backoff retries to handle API request failures can avoid the avalanche effect caused by concentrated retries.70% similarUnverifiedAfter stream resumability is removed, when a long streaming response encounters a network interruption the client can only re-issue the complete request, potentially wasting compute resources and causing inconsistency between the two results.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1051983API
curl https://kongchang.com/api/v1/knowledge/claims/1051983MCP
get_claim(id=1051983)