Mathematical Reasoning on GSM8K (Pass@1, Pass@2)
95.72Pass@1Phi-4-mini + Mistral3-3B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Phi-4-mini + Mistral3-3BTraining Framework=CORE, Model Configuration=Team2026.01 | 95.72 | 96.72 | |
| Phi-4-mini-reasoningTraining Framework=CORE, Model Configuration=Single2026.01 | 90.14 | 91.1 | |
| Mistral3-3B-ReasoningTraining Framework=CORE, Model Configuration=Single2026.01 | 89.25 | 90.23 | |
| Phi-4-mini + Mistral3-3B + OracleTraining Framework=SD-E2, Model Configuration=Oracle2026.01 | 83.25 | 85.5 | |
| Phi-4-mini-reasoningTraining Framework=SD-E2, Model Configuration=Single2026.01 | 80.6 | 84.35 | |
| Mistral3-3B-ReasoningTraining Framework=SD-E2, Model Configuration=Single2026.01 | 79.1 | 83.4 | |
| Phi-4-mini + Mistral3-3B + OracleTraining Framework=Base, Model Configuration=Oracle2026.01 | 74.85 | 80.2 | |
| Phi-4-mini-reasoningTraining Framework=Base, Model Configuration=Single2026.01 | 72.4 | 78.1 | |
| Mistral3-3B-ReasoningTraining Framework=Base, Model Configuration=Single2026.01 | 70.95 | 76.7 |