List your company

Mathematical Reasoning on AMC23 (Avg@16, Avg)

90.62Avg@16OISD
-2.335221.797445.9370.0626May 27, 2026

Evaluation results

MethodLinks
OISDBackbone=Qwen3-8B2026.0590.6266.72
BuPOBackbone=Qwen3-8B2026.0589.2266.36
GRPOBackbone=Qwen3-8B2026.0585.9464.23
OISDBackbone=Qwen3-4B2026.0584.3864.75
BuPOBackbone=Qwen3-4B2026.0581.0958.51
RLOOBackbone=Qwen3-4B2026.0577.6654
PPOBackbone=Qwen3-4B2026.0577.0355.22
GRPOBackbone=Qwen3-4B2026.0576.8855.08
VanillaBackbone=Qwen3-4B2026.0567.6647.44
VanillaBackbone=Qwen3-8B2026.0567.3448.49
Reinforce++Backbone=Qwen3-4B2026.0563.4445.03
OISDBackbone=OctoThinker-8B-Long-Base2026.0539.3830.09
BuPOBackbone=OctoThinker-8B-Long-Base2026.0537.6627.79
GRPOBackbone=OctoThinker-8B-Long-Base2026.0534.8424.11
OISDBackbone=OctoThinker-3B-Long-Base2026.0527.8119.91
GRPOBackbone=OctoThinker-3B-Long-Base2026.0527.518.58
BuPOBackbone=OctoThinker-3B-Long-Base2026.0527.519.59
VanillaBackbone=OctoThinker-8B-Long-Base2026.054.533.75
VanillaBackbone=OctoThinker-3B-Long-Base2026.051.241.68