LLM Serving on Evol-CodeAlpaca (Average Latency)
154.52Average Latency (ms/token)Oracle SJF
Evaluation Results
| Method | Links | |
|---|---|---|
| Oracle SJFModel=DeepSeek-R12025.09 | 154.52 | |
| Oracle SJFModel=Llama2025.09 | 155.97 | |
| PARSModel=Llama2025.09 | 171.62 | |
| PARSModel=DeepSeek-R12025.09 | 171.69 | |
| Cross-Model PARSModel=Llama2025.09 | 182.05 | |
| Listwise SJFModel=Llama2025.09 | 182.4 | |
| Pointwise SJFModel=Llama2025.09 | 184.75 | |
| Listwise SJFModel=DeepSeek-R12025.09 | 200.65 | |
| Cross-Model PARSModel=DeepSeek-R12025.09 | 201.04 | |
| Pointwise SJFModel=DeepSeek-R12025.09 | 203.79 | |
| FCFSModel=Llama2025.09 | 384.18 | |
| FCFSModel=DeepSeek-R12025.09 | 384.28 |