List your company

Mathematical Reasoning on GSM8K (Pass@1, Pass@2)

95.72Pass@1Phi-4-mini + Mistral3-3B
69.959276.647183.33590.0229Jan 29, 2026

Evaluation results

MethodLinks
Phi-4-mini + Mistral3-3BTraining Framework=CORE, Model Configuration=Team2026.0195.7296.72
Phi-4-mini-reasoningTraining Framework=CORE, Model Configuration=Single2026.0190.1491.1
Mistral3-3B-ReasoningTraining Framework=CORE, Model Configuration=Single2026.0189.2590.23
Phi-4-mini + Mistral3-3B + OracleTraining Framework=SD-E2, Model Configuration=Oracle2026.0183.2585.5
Phi-4-mini-reasoningTraining Framework=SD-E2, Model Configuration=Single2026.0180.684.35
Mistral3-3B-ReasoningTraining Framework=SD-E2, Model Configuration=Single2026.0179.183.4
Phi-4-mini + Mistral3-3B + OracleTraining Framework=Base, Model Configuration=Oracle2026.0174.8580.2
Phi-4-mini-reasoningTraining Framework=Base, Model Configuration=Single2026.0172.478.1
Mistral3-3B-ReasoningTraining Framework=Base, Model Configuration=Single2026.0170.9576.7