Unverified50% confidenceFactExact time
Appending messages always works, and modifications to effort, max_tokens, tool_choice, metadata, and cache_control markers likewise work, and server-side compression also does not trigger the problem.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/13/2026
First Seen
Valid until: 1/11/2027
Sources
Related Entities
Related Claims
UnverifiedTool-calling tasks place greater demands on pattern matching and formatted output, and these capabilities are relatively well preserved under quantization compression65% similarUnverified节点间通过话题进行异步消息发布/订阅，适合持续性数据流；通过服务进行同步请求/响应，适合一次性任务63% similarUnverified要最大化缓存命中率，需将稳定不变的内容（系统提示、文件内容）置于Prompt前缀，动态内容追加在末尾63% similarUnverifiedOne API supports multi-channel load balancing, failover, secondary key distribution, usage tracking, and quota control.62% similarUnverifiedPrompt Caching is transparent to the caller: the request still needs to be sent, the network round-trip still exists, the model still runs the remaining inference, and latency remains in the range of seconds.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1022210API
curl https://kongchang.com/api/v1/knowledge/claims/1022210MCP
get_claim(id=1022210)