ResearchBenchmarksLong-context Memory Evaluation on BEHEMOTH LongMemEval (out-of-distribution)Follow63.07AccuracyCluE18.745230.252641.7653.2674Apr 13, 2026Evaluation ResultsMethodMethodLinksAccuracyCluE2026.0463.07MemEvolve2026.0456.82Simple2026.0446.02GEPA2026.0435.06ACE2026.0429.71No Memory2026.0420.45