Mathematical Reasoning on GSM8K (Accuracy Only)
71.87AccuracyTask Arithmetic
Evaluation results
| Method | Links | |
|---|---|---|
| Task ArithmeticParent Model Pair=RL2 & RL3, Parent Model Details=RLinf-math-1.5B & E1-Math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 71.87 | |
| TIES-MergingParent Model Pair=SFT1 & SFT2, Parent Model Details=lul-sft & MiniMath, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 70.5 | |
| SAR-MergingParent Model Pair=RL2 & RL3, Parent Model Details=RLinf-math-1.5B & E1-Math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 70.46 | |
| Linear-MergingParent Model Pair=SFT1 & SFT2, Parent Model Details=lul-sft & MiniMath, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 70.12 | |
| DARE-MergingParent Model Pair=SFT1 & SFT2, Parent Model Details=lul-sft & MiniMath, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 69.74 | |
| SAR-MergingParent Model Pair=RL1 & RL2, Parent Model Details=DeepScaleR-1.5B-Preview & RLinf-math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 68.34 | |
| Naive Avg.Parent Model Pair=SFT1 & SFT2, Parent Model Details=lul-sft & MiniMath, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 68.34 | |
| Naive Avg.Parent Model Pair=RL2 & RL3, Parent Model Details=RLinf-math-1.5B & E1-Math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 67.32 | |
| SAR-MergingParent Model Pair=RL1 & RL3, Parent Model Details=DeepScaleR-1.5B-Preview & E1-Math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 66.75 | |
| Task ArithmeticParent Model Pair=RL1 & RL2, Parent Model Details=DeepScaleR-1.5B-Preview & RLinf-math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 65.42 | |
| TIES-MergingParent Model Pair=RL1 & RL2, Parent Model Details=DeepScaleR-1.5B-Preview & RLinf-math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 65.39 | |
| RAM-MergingParent Model Pair=RL2 & RL3, Parent Model Details=RLinf-math-1.5B & E1-Math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 65.2 | |
| Naive Avg.Parent Model Pair=RL1 & RL3, Parent Model Details=DeepScaleR-1.5B-Preview & E1-Math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 63.03 | |
| Naive Avg.Parent Model Pair=RL1 & RL2, Parent Model Details=DeepScaleR-1.5B-Preview & RLinf-math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 62.88 | |
| TIES-MergingParent Model Pair=RL2 & RL3, Parent Model Details=RLinf-math-1.5B & E1-Math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 61.94 | |
| RAM-MergingParent Model Pair=RL1 & RL2, Parent Model Details=DeepScaleR-1.5B-Preview & RLinf-math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 60.8 | |
| DARE-MergingParent Model Pair=RL1 & RL2, Parent Model Details=DeepScaleR-1.5B-Preview & RLinf-math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 60.57 | |
| Task ArithmeticParent Model Pair=SFT1 & SFT2, Parent Model Details=lul-sft & MiniMath, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 60.27 | |
| DARE-MergingParent Model Pair=RL2 & RL3, Parent Model Details=RLinf-math-1.5B & E1-Math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 58.75 | |
| DARE-MergingParent Model Pair=RL1 & RL3, Parent Model Details=DeepScaleR-1.5B-Preview & E1-Math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 57.71 | |
| Task ArithmeticParent Model Pair=RL1 & RL3, Parent Model Details=DeepScaleR-1.5B-Preview & E1-Math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 56.78 | |
| TIES-MergingParent Model Pair=RL1 & RL3, Parent Model Details=DeepScaleR-1.5B-Preview & E1-Math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 56.33 | |
| Linear-MergingParent Model Pair=RL1 & RL2, Parent Model Details=DeepScaleR-1.5B-Preview & RLinf-math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 55.64 | |
| Linear-MergingParent Model Pair=RL1 & RL3, Parent Model Details=DeepScaleR-1.5B-Preview & E1-Math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 55.64 | |
| Linear-MergingParent Model Pair=RL2 & RL3, Parent Model Details=RLinf-math-1.5B & E1-Math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 55.42 | |
| RAM-MergingParent Model Pair=RL1 & RL3, Parent Model Details=DeepScaleR-1.5B-Preview & E1-Math-1.5B, Model Scale=1.5B, Evaluation Protocol=zero-shot2026.06 | 54.43 |