List your company

Mathematical Reasoning on MATH-500 (maj@16, pass@16)

97.2Maj@16PPC
53.5264.8676.287.54May 28, 2026

Evaluation results

MethodLinks
PPCBackbone=Qwen3-4B2026.0597.299.4
GRPOBackbone=Qwen3-4B2026.0596.698.8
Prompt-OnlyBackbone=Qwen3-4B2026.0596.498.2
Plan-TuningBackbone=Qwen3-4B2026.0596.298.8
BaseBackbone=Qwen3-4B2026.059698.2
PTA-GRPOBackbone=Qwen3-4B2026.0595.898.6
PPCBackbone=Qwen2.5-Math-7B2026.0588.695
PTA-GRPOBackbone=Qwen2.5-Math-7B2026.0587.892.8
Prompt-OnlyBackbone=Qwen2.5-Math-7B2026.0587.691.8
Plan-TuningBackbone=Qwen2.5-Math-7B2026.0587.693.6
BaseBackbone=Qwen2.5-Math-7B2026.058790.8
GRPOBackbone=Qwen2.5-Math-7B2026.0586.893.4
PPCBackbone=Qwen2.5-7B2026.0584.892.2
Plan-TuningBackbone=Qwen2.5-7B2026.058488.8
GRPOBackbone=Qwen2.5-7B2026.0583.890.8
PTA-GRPOBackbone=Qwen2.5-7B2026.0581.489.4
BaseBackbone=Qwen2.5-7B2026.0580.689
Prompt-OnlyBackbone=Qwen2.5-7B2026.0580.490
PPCBackbone=Llama3.1-8B2026.0565.885
PTA-GRPOBackbone=Llama3.1-8B2026.0560.682.8
Plan-TuningBackbone=Llama3.1-8B2026.0560.283
GRPOBackbone=Llama3.1-8B2026.0559.683.4
Prompt-OnlyBackbone=Llama3.1-8B2026.055880
BaseBackbone=Llama3.1-8B2026.0555.282