Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Long-context Memory on LOCOMO

48.88F1 Score

TreeMem

7.373618.149328.92539.7007Apr 2, 2026Apr 7, 2026Apr 13, 2026Apr 19, 2026Apr 24, 2026Apr 30, 2026May 6, 2026
Updated 27d ago

Evaluation Results

MethodLinks
2026.05
48.8843.38
2026.05
45.7739.42
2026.04
45.5638.06
2026.04
45.238.52
2026.04
43.8737.3
2026.05
43.1436.14
2026.04
42.8836.09
2026.04
42.7935.16
2026.04
42.5734.75
2026.04
39.6332.62
2026.04
38.0331.73
2026.04
37.4530.46
2026.04
37.0530.3
2026.04
35.2128.4
2026.04
34.8528.49
2026.04
33.8227.42
2026.04
33.3926.24
2026.04
32.3225.89
2026.04
32.2125.56
2026.05
30.6123.55
2026.04
29.6823.95
2026.04
28.3423.11
2026.05
26.0821.78
2026.04
23.4918.19
2026.05
8.977.27