List your company

Mathematical Reasoning on AIME 26 (Avg@16, Pass@16)

36.25Avg@16CAST
1.586810.585919.58528.5841May 29, 2026

Evaluation results

MethodLinks
CASTModel Scale=Qwen3-4B2026.0536.2570
CASTModel Scale=Qwen3-8B2026.0532.2966.7
GRPOModel Scale=Qwen3-4B2026.052053.3
RLRTModel Scale=Qwen3-4B2026.052063.3
BaseModel Scale=Qwen3-4B2026.0517.7150
RLSDModel Scale=Qwen3-4B2026.0517.7140
RLRTModel Scale=Qwen3-8B2026.0517.7153.3
CASTModel Scale=Qwen3-1.7B2026.0517.2936.7
GRPO+OPSDModel Scale=Qwen3-4B2026.0517.2950
RLSDModel Scale=Qwen3-8B2026.0517.2950
GRPO+OPSDModel Scale=Qwen3-8B2026.0516.6746.7
GRPOModel Scale=Qwen3-8B2026.0515.2140
BaseModel Scale=Qwen3-8B2026.0514.5843.3
OPSDModel Scale=Qwen3-8B2026.0513.7540
RLSDModel Scale=Qwen3-1.7B2026.0511.0433.3
GRPOModel Scale=Qwen3-1.7B2026.051030
BaseModel Scale=Qwen3-1.7B2026.059.3823.3
RLRTModel Scale=Qwen3-1.7B2026.059.3826.7
OPSDModel Scale=Qwen3-4B2026.059.1713.3
GRPO+OPSDModel Scale=Qwen3-1.7B2026.057.523.3
OPSDModel Scale=Qwen3-1.7B2026.052.9213.3