List your company

Mathematical Reasoning on AIME24 (Avg@32, Average)

54.06Avg@32BuPO
-2.162412.433827.0341.6262Dec 22, 2025

Evaluation results

MethodLinks
BuPOBackbone=Qwen3-8B2025.1254.0666.36
GRPOBackbone=Qwen3-8B2025.1249.4864.23
RLOOBackbone=Qwen3-8B2025.1246.6763.36
Reinforce++Backbone=Qwen3-8B2025.1241.7760.41
PPOBackbone=Qwen3-8B2025.1237.8158.41
BuPOBackbone=Qwen3-4B2025.1236.8858.51
PPOBackbone=Qwen3-4B2025.1232.655.22
GRPOBackbone=Qwen3-4B2025.1232.1955.08
RLOOBackbone=Qwen3-4B2025.1230.8354
VanillaBackbone=Qwen3-8B2025.1226.9848.49
VanillaBackbone=Qwen3-4B2025.1223.247.44
Reinforce++Backbone=Qwen3-4B2025.1217.445.03
Reinforce++Backbone=Llama-OctoThinker-8B-Base2025.127.7226.43
BuPOBackbone=Llama-OctoThinker-8B-Base2025.124.6927.79
RLOOBackbone=Llama-OctoThinker-8B-Base2025.123.5422.18
GRPOBackbone=Llama-OctoThinker-8B-Base2025.122.524.11
RLOOBackbone=Llama-OctoThinker-3B-Base2025.122.1917.84
PPOBackbone=Llama-OctoThinker-8B-Base2025.121.5622.82
PPOBackbone=Llama-OctoThinker-3B-Base2025.121.0416.69
GRPOBackbone=Llama-OctoThinker-3B-Base2025.120.6318.58
BuPOBackbone=Llama-OctoThinker-3B-Base2025.120.6319.59
VanillaBackbone=Llama-OctoThinker-8B-Base2025.120.523.75
VanillaBackbone=Llama-OctoThinker-3B-Base2025.120.211.68
Reinforce++Backbone=Llama-OctoThinker-3B-Base2025.1205.27