List your company

Mathematical Reasoning on AIME 24 (Avg@32)

27.92Avg@32GSPO
0.7247.784514.84521.9055Feb 9, 2026

Evaluation results

MethodLinks
GSPOBackbone=Qwen3-8B-Base2026.0227.92
GRPO + CBS*Backbone=Qwen3-8B-Base2026.0226.46
GSPO + CBS*Backbone=Qwen3-8B-Base2026.0225.1
DAPO + CBS*Backbone=Qwen3-8B-Base2026.0224.38
GRPO + CBSBackbone=Qwen3-8B-Base2026.0223.65
DAPO + CBSBackbone=Qwen3-8B-Base2026.0223.54
GSPO + CBSBackbone=Qwen3-8B-Base2026.0223.02
DAPOBackbone=Qwen3-8B-Base2026.0222.29
GSPO + CBSBackbone=Qwen3-4B-Base2026.0221.46
GSPO + CBS*Backbone=Qwen3-4B-Base2026.0221.25
GRPOBackbone=Qwen3-4B-Base2026.0221.04
DAPO + CBSBackbone=Qwen3-4B-Base2026.0220.62
GRPO + CBS*Backbone=Qwen3-4B-Base2026.0220.52
GRPOBackbone=Qwen3-8B-Base2026.0220.52
GRPO + CBSBackbone=Qwen3-4B-Base2026.0220.31
DAPO + CBS*Backbone=Qwen3-4B-Base2026.0218.96
GSPOBackbone=Qwen3-4B-Base2026.0218.44
DAPOBackbone=Qwen3-4B-Base2026.0216.88
GSPO + CBS*Backbone=Qwen3-1.7B-Base2026.0212.5
GRPO + CBS*Backbone=Qwen3-1.7B-Base2026.0211.98
GSPO + CBSBackbone=Qwen3-1.7B-Base2026.0211.35
GSPOBackbone=Qwen3-1.7B-Base2026.0210.73
DAPO + CBS*Backbone=Qwen3-1.7B-Base2026.0210.52
GRPOBackbone=Qwen3-1.7B-Base2026.029.58
DAPO + CBSBackbone=Qwen3-1.7B-Base2026.028.96
Base ModelBackbone=Qwen3-8B-Base2026.028.33
GRPO + CBSBackbone=Qwen3-1.7B-Base2026.028.12
DAPOBackbone=Qwen3-1.7B-Base2026.026.35
Base ModelBackbone=Qwen3-4B-Base2026.025.1
Base ModelBackbone=Qwen3-1.7B-Base2026.021.77