Expired50% confidenceFactExact time
Short-term memory is implemented by concatenating conversation history in the prompt, limited by the model's context length, which typically ranges from 4K to 200K tokens.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/23/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/21/2026(expired)
Sources
Related Claims
UnverifiedPersistent memory is typically divided into three levels: short-term working memory, episodic memory, and semantic memory71% similarUnverified对于单用户交互式场景推理速度达到5-10 token/s通常就能提供流畅体验，而代码生成、长文档摘要或多轮Agent调用往往需要30 token/s以上71% similarVerifiedLangGraph supports two storage modes: short-term memory (single-session context) and long-term memory (cross-session persistent storage).70% similarVerifiedAgent memory is divided into four levels: short-term memory, long-term memory, entity memory, and procedural memory.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/597081API
curl https://kongchang.com/api/v1/knowledge/claims/597081MCP
get_claim(id=597081)