6x RTX 4090
10 records across 1 models (quants: UD-Q4_K_XL).
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 6x RTX 4090 | Qwen3.8-Flash-Next | UD-Q4_K_XL | llama.cpp (multigpu fork) | 57.5 | – | – | – | 51200 | 2026-10-06 | community | – | source |
| 6x RTX 4090 | Qwen3.8-Flash-Next | UD-Q4_K_XL | llama.cpp (multigpu fork) | 48.7 | – | – | – | 256000 | 2026-10-06 | community | – | source |
| 6x RTX 4090 | Qwen3.8-Flash-Next | UD-Q4_K_XL | llama.cpp (CUDA) | 46.2 | – | – | – | 51200 | 2026-10-06 | community | – | source |
| 6x RTX 4090 | Qwen3.8-Flash-Next | UD-Q4_K_XL | llama.cpp (multigpu fork) | 30.8 | – | – | – | 256000 | 2026-10-06 | community | – | source |
| 6x RTX 4090 | Qwen3.8-Flash-Next | UD-Q4_K_XL | llama.cpp (CUDA) | 21.0 | – | – | – | 256000 | 2026-10-06 | community | – | source |
| 6x RTX 4090 | Qwen3.8-Flash-Next | UD-Q4_K_XL | llama.cpp (CUDA) | – | 1728 | – | – | 51200 | 2026-10-06 | community | – | source |
| 6x RTX 4090 | Qwen3.8-Flash-Next | UD-Q4_K_XL | llama.cpp (multigpu fork) | – | 7050 | – | – | 51200 | 2026-10-06 | community | – | source |
| 6x RTX 4090 | Qwen3.8-Flash-Next | UD-Q4_K_XL | llama.cpp (CUDA) | – | 744 | – | – | 256000 | 2026-10-06 | community | – | source |
| 6x RTX 4090 | Qwen3.8-Flash-Next | UD-Q4_K_XL | llama.cpp (multigpu fork) | – | 7403 | – | – | 256000 | 2026-10-06 | community | – | source |
| 6x RTX 4090 | Qwen3.8-Flash-Next | UD-Q4_K_XL | llama.cpp (multigpu fork) | – | 7241 | – | – | 256000 | 2026-10-06 | community | – | source |
All rows cite their source. Provenance: community 40018 sourced 1085