List your company

Mathematical Reasoning on AIME24 (ACC, mean@8)

35AccuracyDiPO
5.963213.501621.0428.5784Apr 15, 2026

Evaluation results

MethodLinks
DiPOBackbone=Qwen3-8B-Base2026.043535
DAPO w/ ELBackbone=Qwen3-8B-Base2026.0433.7533.75
GRPOBackbone=Qwen3-8B-Base2026.0431.6731.67
CDEBackbone=Qwen3-8B-Base2026.0431.6731.67
DAPOBackbone=Qwen3-8B-Base2026.0430.0830.08
DiPOBackbone=Qwen3-4B-Base2026.0429.1729.17
GRPOBackbone=Qwen3-4B-Base2026.0426.6726.67
DAPO w/ ELBackbone=Qwen3-4B-Base2026.0426.6726.67
CDEBackbone=Qwen3-4B-Base2026.0426.6726.67
DAPOBackbone=Qwen3-4B-Base2026.0426.2526.25
DiPOBackbone=Qwen2.5-7B2026.0422.9222.92
GRPOBackbone=Qwen2.5-7B2026.0420.4220.42
DAPOBackbone=Qwen2.5-7B2026.0420.4220.42
DAPO w/ ELBackbone=Qwen2.5-7B2026.042020
CDEBackbone=Qwen2.5-7B2026.042020
Base modelBackbone=Qwen3-4B-Base2026.049.589.58
Base modelBackbone=Qwen3-8B-Base2026.048.338.33
Base modelBackbone=Qwen2.5-7B2026.047.087.08