Unverified50% confidenceFactExact time
The corresponding fix commit is 3d6ac0a, which maps incomplete_details.reason to the finish reason for generateText calls in Responses mode.
1
Sources
50%
Confidence
Long-term
Relevance
11/2/2026
First Seen
Sources
Related Entities
Related Claims
Unverifiedv1.4.0 fixes an abnormal merge logic for custom-code in the messages resource.66% similarUnverifiedincomplete_details.reason is a field natively returned by the xAI Responses API and may contain reason strings such as max_output_tokens and content_filter.63% similarUnverifiedResolving text rendering issues typically requires introducing a glyph encoder or OCR auxiliary module that injects the visual form of characters as an additional condition into the generation process.61% similarUnverifiedThe fix was to split the Build prompt into a cached portion containing no per-message changing content and a separately sent per-turn tail block, without changing the prompt text itself.60% similarUnverifiedrerere only compares text and does not encode semantics, so a replayed resolution may be wrong for code that has changed elsewhere.59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1128144API
curl https://kongchang.com/api/v1/knowledge/claims/1128144MCP
get_claim(id=1128144)