| Benchmark | Model | Rank | Results |
|---|---|---|---|
| arithmetic-reasoning-on-gsm8k | CharacterFlywheel V7 (Llama 3.1 70B) | – | Accuracy: 92.3 |
| code-generation-on-humaneval | CharacterFlywheel V7 (Llama 3.1 70B) | – | Pass@1: 77.4 |
| code-generation-on-mbpp | CharacterFlywheel V7 (Llama 3.1 70B) | – | Accuracy: 66.6 |
| math-word-problem-solving-on-math | CharacterFlywheel V7 (Llama 3.1 70B) | – | Accuracy: 50.5 |