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

Long-context language understanding on LongBench (Aggregate Score Only)

49.8Aggregate Score

FullKV

24.299230.919637.5444.1604Jun 23, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
49.8
2026.06
49.08
2026.06
48.69
2026.06
48.57
2026.06
48.52
2026.06
48.42
2026.06
48.24
2026.06
48.19
2026.06
48.18
2026.06
47.97
2026.06
47.94
2026.06
47.82
2026.06
47.82
2026.06
47.78
2026.06
47.7
2026.06
47.65
2026.06
47.51
2026.06
47.5
2026.06
47.34
2026.06
47.33
2026.06
47.05
2026.06
46.98
2026.06
46.96
2026.06
46.71
2026.06
46.66
2026.06
46.48
2026.06
46.41
2026.06
46.38
2026.06
46.3
2026.06
45.96
2026.06
45.43
2026.06
45.37
2026.06
45.21
2026.06
44.73
2026.06
44.73
2026.06
44.7
2026.06
44.49
2026.06
44.45
2026.06
44.36
2026.06
44.21
2026.06
44.11
2026.06
44.1
2026.06
44.02
2026.06
43.77
2026.06
43.76
2026.06
43.75
2026.06
43.68
2026.06
43.36
2026.06
43.3
2026.06
43.06
2026.06
42.71
2026.06
42.11
2026.06
36.95
2026.06
34.73
2026.06
33.92
2026.06
31.22
2026.06
29.84
2026.06
29.62
2026.06
25.86
2026.06
25.28