| Benchmark | Model | Rank | Results |
|---|---|---|---|
| math-word-problem-solving-on-math | DeepSeekMATH-RL-7B (w/ code, greedy decoding) | #57 | Accuracy: 58.8Parameters (Billions): 7 |
| math-word-problem-solving-on-math | DeepSeekMATH-RL-7B (greedy decoding) | #80 | Accuracy: 51.7Parameters (Billions): 7 |
| arithmetic-reasoning-on-gsm8k | DeepSeekMATH-RL-7B | #92 | Accuracy: 88.2Parameters (Billion): 7 |