4× NVIDIA CMP 170HX
11 records across 1 models (quants: EXL3 2/3 bpw mixed).
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 4× NVIDIA CMP 170HX | DeepSeek-V4.1-Flash 764B | EXL3 2/3 bpw mixed | vLLM | 252.1 | – | – | – | – | 2026-09-29 | community | – | source |
| 4× NVIDIA CMP 170HX | DeepSeek-V4.1-Flash 764B | EXL3 2/3 bpw mixed | vLLM | 200.6 | – | – | – | – | 2026-09-29 | community | – | source |
| 4× NVIDIA CMP 170HX | DeepSeek-V4.1-Flash 764B | EXL3 2/3 bpw mixed | vLLM | 199.4 | – | – | – | – | 2026-09-29 | community | – | source |
| 4× NVIDIA CMP 170HX | DeepSeek-V4.1-Flash 764B | EXL3 2/3 bpw mixed | vLLM | 60.7 | – | – | – | – | 2026-09-29 | community | – | source |
| 4× NVIDIA CMP 170HX | DeepSeek-V4.1-Flash 764B | EXL3 2/3 bpw mixed | vLLM | 58.3 | – | – | – | – | 2026-09-29 | community | – | source |
| 4× NVIDIA CMP 170HX | DeepSeek-V4.1-Flash 764B | EXL3 2/3 bpw mixed | vLLM | 57.0 | – | – | – | – | 2026-09-29 | community | – | source |
| 4× NVIDIA CMP 170HX | DeepSeek-V4.1-Flash 764B | EXL3 2/3 bpw mixed | vLLM | 56.7 | – | – | – | – | 2026-09-29 | community | – | source |
| 4× NVIDIA CMP 170HX | DeepSeek-V4.1-Flash 764B | EXL3 2/3 bpw mixed | vLLM | 54.7 | – | – | – | – | 2026-09-29 | community | – | source |
| 4× NVIDIA CMP 170HX | DeepSeek-V4.1-Flash 764B | EXL3 2/3 bpw mixed | vLLM | 46.7 | – | – | – | – | 2026-09-29 | community | – | source |
| 4× NVIDIA CMP 170HX | DeepSeek-V4.1-Flash 764B | EXL3 2/3 bpw mixed | vLLM | 40 | – | – | – | – | 2026-09-29 | community | – | source |
| 4× NVIDIA CMP 170HX | DeepSeek-V4.1-Flash 764B | EXL3 2/3 bpw mixed | vLLM | 35.8 | – | – | – | – | 2026-09-29 | community | – | source |
All rows cite their source. Provenance: community 40018 sourced 1085