Unverified60% confidenceFactTime unknown
根据LongMEM Eval基准测试，AgentMemory的检索准确率达到95.2%，比同类项目Mem0高出27个百分点
2
Sources
60%
Confidence
Medium-term (~90 days)
Relevance
6/3/2026
First Seen
Valid until: 9/1/2026
Sources
AgentMemory：给AI编码助手装上永久记忆的开源方案
bilibili锋芒AI
Related Entities
Related Claims
UnverifiedMemPalace在LongMemEval基准测试上，不使用LLM的纯检索方案前5召回率达到96.6%69% similarUnverifiedAI意图识别准确率一般在85%-95%之间67% similarUnverified移除歧义后所有模型准确率提升26.8至40.2个百分点，Gemini甚至达到81%，表明歧义是深度搜索代理最大性能瓶颈之一66% similarUnverified美团LongCat推出Lohosearch搜索智能体基准，11个前沿模型测试中最佳得分仅34.74%，远低于人类约90%的表现65% similarUnverifiedDeepSuite基准测试的误报率为44.9%，漏报率为1.2%65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/39397API
curl https://kongchang.com/api/v1/knowledge/claims/39397MCP
get_claim(id=39397)