| 量子化 | Size | OS | CPU | RAM | RAM | GPU | Window | token/s | software | |
|---|---|---|---|---|---|---|---|---|---|---|
| UD_Q4_K_XL | 125 GB | Ubuntu 24.04 | Ryzen 9 3950X | DDR4-3200 | 128GB | GeForce RTX 4060Ti 16GB | 4096 | 4.8 tps | llama.cpp b8123 | |
| UD_Q3_K_XL | 97 GB | Ubuntu 24.04 | Ryzen 9 3950X | DDR4-3200 | 128GB | GeForce RTX 4060Ti 16GB | 4096 | 6.2 tps | llama.cpp b8123 | |
| UD_Q3_K_XL | 97 GB | Windows11 | Ryzen 7 9700X | DDR5-5600 | 128GB | GeForce RTX 5060Ti 16GB | 4096 | 6.81 tps | llama.cpp b8123 | |
| UD_Q3_K_XL | 97 GB | Windows11 | Ryzen 7 9700X | DDR5-5600 | 128GB | GeForce RTX 5060Ti 16GB | 4096 | 6.59 tps | LMStudio | cpu moe 85 |