Long-context Memory Retrieval and Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
48.01Single-Hop Retrieval F1
20
Apr 3, 2026
38.21F1 Score
20
Mar 5, 2026
27.44F1 Score
20
Mar 5, 2026
24.95F1 Score
20
Mar 5, 2026
51.68F1 Score
20
Mar 5, 2026
24.49F1 Score
20
Mar 5, 2026
23.75F1 Score
20
Mar 5, 2026
26.45F1 Score
20
Mar 5, 2026
49.58F1 Score
20
Mar 5, 2026
47.26F1 Score
20
Mar 5, 2026
46.39F1 Score
20
Mar 5, 2026
89.7KU L-J Score
10
Jun 11, 2026
78.1Single-hop L-J Score
10
Jun 11, 2026