List your company

Mathematical Reasoning on AMC23 (Avg@32)

90Avg@32DSRL
67.411273.275679.1485.0044Apr 15, 2026

Evaluation results

MethodLinks
DSRLBackbone=Qwen3-8B2026.0490
PPOBackbone=Qwen3-8B2026.0489.69
DSRLBackbone=Qwen3-4B2026.0489.22
DAPOBackbone=Qwen3-8B2026.0488.52
DAPOBackbone=Qwen3-4B2026.0488.2
GRPOBackbone=Qwen3-8B2026.0488.05
PPOBackbone=Qwen3-4B2026.0487.89
GRPOBackbone=Qwen3-4B2026.0487.81
Dr.GRPOBackbone=Qwen3-8B2026.0487.66
Dr.GRPOBackbone=Qwen3-4B2026.0487.58
RLOOBackbone=Qwen3-8B2026.0487.27
RLOOBackbone=Qwen3-4B2026.0485
Reinforce++Backbone=Qwen3-8B2026.0484.14
Reinforce++Backbone=Qwen3-4B2026.0479.37
VanillaBackbone=Qwen3-8B2026.0469.53
VanillaBackbone=Qwen3-4B2026.0468.28