ResearchBenchmarksMemory-intensive task on LoCoMoFollow41.28F1 ScoreNapMem31.004833.672436.3439.0076Jul 7, 2026Evaluation ResultsMethodMethodLinksF1 ScoreLLM-Judge Accuracy (L-J)NapMemModel Size=9B, Optimiz...Model Size=9B, Optimization=w/ RL2026.0741.2859.92Mem02026.0741.1960.86NapMemModel Size=397BModel Size=397B2026.0738.3154.42AgeMem2026.073845.02Zep2026.0736.0952.21NapMemModel Size=122BModel Size=122B2026.0735.3951.69MemOS2026.0735.2255.82MemoryOS2026.0731.444.28