| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Instruction Following | Vicuna & WizardLM Average | Win Rate vs ChatGPT59.7 | 9 | |
| Instruction Following | Vicuna & WizardLM Urdu | Win Rate (vs ChatGPT)71.1 | 9 | |
| Instruction Following | Vicuna & WizardLM Tamil ta | Win Rate (vs ChatGPT)76.8 | 9 | |
| Instruction Following | Vicuna & WizardLM Swahili | Win Rate (vs ChatGPT)54.4 | 9 | |
| Instruction Following | Vicuna & WizardLM Hindi | Win Rate (vs ChatGPT)65.8 | 9 | |
| Instruction Following | Vicuna & WizardLM Bengali bn | Win Rate (vs ChatGPT)68.8 | 9 | |
| Instruction Following | Vicuna & WizardLM Vietnamese / vi | Win Rate (vs ChatGPT)57 | 9 | |
| Instruction Following | Vicuna & WizardLM Turkish | Win Rate (vs ChatGPT)53.7 | 9 | |
| Instruction Following | Vicuna & WizardLM Thai | Win Rate (vs ChatGPT)53 | 9 | |
| Instruction Following | Vicuna & WizardLM Indonesian | Win Rate (vs ChatGPT)50.3 | 9 | |
| Instruction Following | Vicuna & WizardLM Finnish fi | Win Rate (vs ChatGPT)47 | 9 |