Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Mathematical Reasoning on AIME 2025 (Accuracy, CLC)
Loading...
60.1
Accuracy
SOLAR
21.568
31.5715
41.575
51.5785
Jun 25, 2026
Accuracy
CLC
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
CLC
SOLAR
Backbone=Qwen3-8B, Fin...
2026.06
60.1
53.02
SFT
Backbone=Qwen3-8B, Fin...
2026.06
58.38
50.13
Qwen3-8B
Backbone=Qwen3-8B, Fin...
2026.06
54.91
45.06
Soft Thinking
Backbone=Qwen3-8B, Fin...
2026.06
54.71
44.22
SOLAR
Backbone=Qwen3-4B, Fin...
2026.06
30.95
24.51
MidAlign
Backbone=Qwen3-4B, Fin...
2026.06
29.05
23.22
SFT
Backbone=Qwen3-4B, Fin...
2026.06
27.9
20.44
AlignX
Backbone=Qwen3-4B, Fin...
2026.06
27.76
20.1
MAPO
Backbone=Qwen3-4B, Fin...
2026.06
25.81
18.94
Soft Thinking
Backbone=Qwen3-4B, Fin...
2026.06
23.33
18.22
Qwen3-4B
Backbone=Qwen3-4B, Fin...
2026.06
23.05
17.61
Feedback
Search any
task
Search any
task