List your company

Mathematical Reasoning on Minerva-Math (maj@16, pass@16)

46.32maj@16PPC
34.089637.264840.4443.6152May 28, 2026

Evaluation results

MethodLinks
PPCBackbone=Qwen3-4B2026.0546.3260.66
PTA-GRPOBackbone=Qwen3-4B2026.0545.2258.09
Plan-TuningBackbone=Qwen3-4B2026.0544.8556.25
PPCBackbone=Qwen2.5-7B2026.0544.8562.87
GRPOBackbone=Qwen3-4B2026.0544.4855.88
Prompt-OnlyBackbone=Qwen3-4B2026.0543.7554.41
PPCBackbone=Qwen2.5-Math-7B2026.0543.3858.09
BaseBackbone=Qwen3-4B2026.0543.0155.15
PTA-GRPOBackbone=Qwen2.5-7B2026.0543.0158.82
GRPOBackbone=Qwen2.5-Math-7B2026.0542.6553.31
Prompt-OnlyBackbone=Qwen2.5-Math-7B2026.0541.9252.21
Plan-TuningBackbone=Qwen2.5-7B2026.0541.5455.88
GRPOBackbone=Qwen2.5-7B2026.0541.1861.76
BaseBackbone=Qwen2.5-Math-7B2026.0540.8151.84
BaseBackbone=Qwen2.5-7B2026.0540.0754.78
Plan-TuningBackbone=Qwen2.5-Math-7B2026.0540.0752.94
Prompt-OnlyBackbone=Qwen2.5-7B2026.0539.7158.09
PTA-GRPOBackbone=Qwen2.5-Math-7B2026.0537.8756.62
PPCBackbone=Llama3.1-8B2026.0537.561.4
PTA-GRPOBackbone=Llama3.1-8B2026.0537.1359.93
GRPOBackbone=Llama3.1-8B2026.0536.459.19
Plan-TuningBackbone=Llama3.1-8B2026.0536.0358.09
BaseBackbone=Llama3.1-8B2026.0535.2960.29
Prompt-OnlyBackbone=Llama3.1-8B2026.0534.5659.56