RTX 5090 ×2
1 records across 1 models (quants: Q4_K_M).
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| RTX 5090 ×2 | Qwen3-Coder-Next | Q4_K_M | llama.cpp | 74 | – | – | – | – | 2026-05 | community | – | source |
All rows cite their source. Provenance: community 40018 sourced 1085