List your company

Mathematical Reasoning on AIME24 (Avg@32, Avg)

54.06AIME24 Avg@32 ScoreBuPO
-1.94412.595527.13541.6745May 27, 2026

Evaluation results

MethodLinks
BuPOBackbone=Qwen3-8B2026.0554.0666.36
OISDBackbone=Qwen3-8B2026.0552.466.72
GRPOBackbone=Qwen3-8B2026.0549.4864.23
OISDBackbone=Qwen3-4B2026.0547.2964.75
BuPOBackbone=Qwen3-4B2026.0536.8858.51
PPOBackbone=Qwen3-4B2026.0532.655.22
GRPOBackbone=Qwen3-4B2026.0532.1955.08
RLOOBackbone=Qwen3-4B2026.0530.8354
VanillaBackbone=Qwen3-8B2026.0526.9848.49
VanillaBackbone=Qwen3-4B2026.0523.247.44
Reinforce++Backbone=Qwen3-4B2026.0517.445.03
OISDBackbone=OctoThinker-8B-Long-Base2026.057.2930.09
BuPOBackbone=OctoThinker-8B-Long-Base2026.054.6927.79
GRPOBackbone=OctoThinker-8B-Long-Base2026.052.524.11
OISDBackbone=OctoThinker-3B-Long-Base2026.051.9819.91
GRPOBackbone=OctoThinker-3B-Long-Base2026.050.6318.58
BuPOBackbone=OctoThinker-3B-Long-Base2026.050.6319.59
VanillaBackbone=OctoThinker-8B-Long-Base2026.050.523.75
VanillaBackbone=OctoThinker-3B-Long-Base2026.050.211.68