Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multilingual Mathematical Reasoning on MGSM (Accuracy, CLC)
Loading...
88.27
Accuracy
SOLAR
73.4396
77.2898
81.14
84.9902
Jun 25, 2026
Accuracy
CLC
Updated 29d ago
Evaluation Results
Method
Method
Links
Accuracy
CLC
SOLAR
Backbone=Qwen3-8B, Fin...
2026.06
88.27
87.28
SFT
Backbone=Qwen3-8B, Fin...
2026.06
86.9
84.71
Qwen3-8B
Backbone=Qwen3-8B, Fin...
2026.06
83.42
80.51
Soft Thinking
Backbone=Qwen3-8B, Fin...
2026.06
83.04
80.01
SOLAR
Backbone=Qwen3-4B, Fin...
2026.06
78.27
75.61
AlignX
Backbone=Qwen3-4B, Fin...
2026.06
77.69
73
MAPO
Backbone=Qwen3-4B, Fin...
2026.06
77.52
73.9
MidAlign
Backbone=Qwen3-4B, Fin...
2026.06
76.45
73.11
SFT
Backbone=Qwen3-4B, Fin...
2026.06
76.25
72.29
Soft Thinking
Backbone=Qwen3-4B, Fin...
2026.06
74.23
70.37
Qwen3-4B
Backbone=Qwen3-4B, Fin...
2026.06
74.01
69.31
Feedback
Search any
task
Search any
task