A100
1recorded run
33.7best tok/s
0sourced
| GPU | Model | Quant | Backend | Ctx | tok/s | prompt tok/s | VRAM GB | Source |
|---|---|---|---|---|---|---|---|---|
| A100 · 40 GB | Qwen3.8-27B | Q8_0 | llama.cpp — | 65536 | 33.7 … | 1088.2 | 30.0 | src |
| GPU | Model | Quant | Backend | Ctx | tok/s | prompt tok/s | VRAM GB | Source |
|---|---|---|---|---|---|---|---|---|
| A100 · 40 GB | Qwen3.8-27B | Q8_0 | llama.cpp — | 65536 | 33.7 … | 1088.2 | 30.0 | src |