ResearchBenchmarksLong-horizon reasoning on LoCoMoFollow73.9Average ScoreHAGE49.14855.5746268.426May 11, 2026Evaluation ResultsMethodMethodLinksAverage ScoreToken Efficiency (Tokens/Query)Inference Latency (s)HAGE2026.0573.93.822.17MAGMA2026.05703.371.72Nemori2026.05593.462.59A-MEM2026.05582.622.26MemoryOS2026.0555.34.7632.68MemSkill2026.0550.10.921.46