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

Long-context Reasoning on LongGenBench 4K

53.18GSM8K Score

Full Cache

7.461619.330831.243.0692May 28, 2026
Updated 5d ago

Evaluation Results

MethodLinks
2026.05
53.1854.2771.5459.63
2026.05
52.8655.1272.8360.27
2026.05
52.0152.8271.9158.91
2026.05
50.9354.8471.9159.23
2026.05
50.3853.8169.6657.95
2026.05
50.3855.6373.6659.89
2026.05
48.6854.7872.4158.62
2026.05
46.253.669.5856.46
2026.05
44.4955.5771.8357.3
2026.05
38.7551.4264.5851.58
2026.05
37.151.863.550.8
2026.05
13.4128.4566.0435.97
2026.05
12.6328.5765.8435.68
2026.05
11.9428.1764.2134.77
2026.05
11.9327.0664.3334.44
2026.05
11.3925.0964.7533.74
2026.05
11.2427.8163.9134.32
2026.05
11.1628.246434.47
2026.05
11.1624.0365.0833.42
2026.05
9.9226.8761.232.66
2026.05
9.9226.8761.2532.68
2026.05
9.2228.5165.4134.38