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

Code Generation on LiveCodeBench v6 (Accuracy, Mean, Drop)

59.06Accuracy

BF16

-2.362413.583829.5345.4762May 18, 2026
Updated 14d ago

Evaluation Results

MethodLinks
2026.05
59.0674.19-
2026.05
58.4878.150.26
2026.05
56.0274.430.24
2026.05
55.5671.99-2.2
2026.05
53.5774.17-0.02
2026.05
52.6378.160.27
2026.05
50.4977.950.06
2026.05
49.1277.89-
2026.05
49.0170.84-
2026.05
48.6675.64-
2026.05
48.5475.14-2.75
2026.05
47.9569.97-0.87
2026.05
46.3269.416-1.42
2026.05
46.273.11-2.53
2026.05
45.3871.864-3.78
2026.05
37.0360.49-17.4
2026.05
21.0556.88-13.96
2026.05
0.5831.74-43.9
2026.05
0.5810.14-60.7
2026.05
0.397.9-66.29
2026.05
01.4-74.24
2026.05
00-75.64
2026.05
00-70.84
2026.05
00-74.19