2x RTX 4090

3 records across 1 models (quants: Q4_0).

hardwaremodelquantbackendtok/spp tok/sWttft sctx/tgdateclassflagssource
2x RTX 4090Qwen3.8-Flash-Next Q4_0llama.cpp (CUDA)64.39321.5–––2026-10-07community–source
2x RTX 4090Qwen3.8-Flash-Next Q4_0llama.cpp (CUDA)60.38402.0–––2026-10-07community–source
2x RTX 4090Qwen3.8-Flash-Next Q4_0llama.cpp (CUDA)34.78440.7–––2026-10-07community–source

All rows cite their source. Provenance: community 40018 sourced 1085

← all hardware