List your company

Mathematical Reasoning on AIME26 (Acc, Think Tok.)

11,381Thinking TokensDRE
10,984.0813,663.2916,342.519,021.71Jun 16, 2026

Evaluation results

MethodLinks
DREBase Model=Qwen3-8B2026.0611,38166.87
DREBackbone=Qwen3-8B2026.0611,38166.87
DREBase Model=Qwen3-30B-A3B-Thinking-25072026.0612,15386.46
DREBackbone=Qwen3-30B-A3B-Thinking-25072026.0612,15386.46
DEERBackbone=Qwen3-8B2026.0614,32459.26
DREBase Model=Qwen3-4B-Thinking-25072026.0614,43179.37
DREBackbone=Qwen3-4B-Thinking-25072026.0614,43179.37
RCPDBackbone=Qwen3-8B2026.0614,85762.5
GRPO + LPBase Model=Qwen3-8B2026.0615,10763.96
GRPOBase Model=Qwen3-30B-A3B-Thinking-25072026.0615,15884.58
GRPOBase Model=Qwen3-8B2026.0615,58165.21
RCPDBackbone=Qwen3-30B-A3B-Thinking-25072026.0615,74185.83
Raw ModelBase Model=Qwen3-8B2026.0615,86266.67
Raw ModelBackbone=Qwen3-8B2026.0615,86266.67
GRPO + LPBase Model=Qwen3-30B-A3B-Thinking-25072026.0616,03985
RCPDBackbone=Qwen3-4B-Thinking-25072026.0616,91676.04
Raw ModelBase Model=Qwen3-30B-A3B-Thinking-25072026.0616,98588.96
Raw ModelBackbone=Qwen3-30B-A3B-Thinking-25072026.0616,98588.96
S-GRPOBase Model=Qwen3-8B2026.0617,70762.5
DEERBackbone=Qwen3-30B-A3B-Thinking-25072026.0617,70786.29
S-GRPOBase Model=Qwen3-30B-A3B-Thinking-25072026.0619,13385
DEERBackbone=Qwen3-4B-Thinking-25072026.0619,21978.52
GRPOBase Model=Qwen3-4B-Thinking-25072026.0620,28476.67
GRPO + LPBase Model=Qwen3-4B-Thinking-25072026.0620,30275.62
Raw ModelBase Model=Qwen3-4B-Thinking-25072026.0620,78977.71
Raw ModelBackbone=Qwen3-4B-Thinking-25072026.0620,78977.71
S-GRPOBase Model=Qwen3-4B-Thinking-25072026.0621,30478.33