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

Long-context Reasoning on LongReason Average across 16k-64k

54.2Accuracy

Full

38.787242.788646.7950.7914Jun 25, 2026
Updated 29d ago

Evaluation Results

MethodLinks
2026.06
54.2
2026.06
52.53
2026.06
51.13
2026.06
51.09
2026.06
50.76
2026.06
50.15
2026.06
49.79
2026.06
49.37
2026.06
48.88
2026.06
48.7
2026.06
48.57
2026.06
47.15
2026.06
46.89
2026.06
46.81
2026.06
46.55
2026.06
46.25
2026.06
45.83
2026.06
45.7
2026.06
45.68
2026.06
44.42
2026.06
43.95
2026.06
43.11
2026.06
43.03
2026.06
42.86
2026.06
42.23
2026.06
41.81
2026.06
41.65
2026.06
41.56
2026.06
41.39
2026.06
40.75
2026.06
40.72
2026.06
40.67
2026.06
39.67
2026.06
39.42
2026.06
39.42
2026.06
39.38