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

Math Problem Solving on GSM8K (solve rate)

90.22Solve Rate

Bifrost

0.249623.607346.96570.3227Feb 5, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.02
90.22
2026.02
87.41
2026.02
86.13
2026.02
81.41
2026.02
80.78
2026.02
78.62
2026.02
76.8
2026.02
76.72
2026.02
74.65
2026.02
70.18
2026.02
62.52
2026.02
62.4
2026.02
55.27
2026.02
48.29
2026.02
44.66
2026.02
42.15
2026.02
40.71
2026.02
40.26
2026.02
39.73
2026.02
36.92
2026.02
33.74
2026.02
23.81
2026.02
21.08
2026.02
15.85
2026.02
14.48
2026.02
9.62
2026.02
3.71