RTX 4070

13 records across 3 models (quants: PTQ1_0, Q4_K).

hardwaremodelquantbackendtok/spp tok/sWttft sctx/tgdateclassflagssource
RTX 4070Qwen3 14B 14BQ4_Kllama.cpp42.52099.9––4096–sourced–source
RTX 4070Qwen3 14B 14BQ4_Kllama.cpp32.71355.8––16384–sourced–source
RTX 4070Qwen3 8B 8BQ4_Kllama.cpp71.23564.1––4096–sourced–source
RTX 4070Qwen3 8B 8BQ4_Kllama.cpp52.12064.3––16384–sourced–source
RTX 4070Qwen3 8B 8BQ4_Kllama.cpp38.11116.7––32768–sourced–source
RTX 4070Ternary-Bonsai-2-27B PTQ1_0llama.cpp (ada-surgery fork)1021173––40962026-10-08community–source
RTX 4070Ternary-Bonsai-2-27B PTQ1_0llama.cpp (ada-surgery fork)831137––163842026-10-08community–source
RTX 4070Ternary-Bonsai-2-27B PTQ1_0llama.cpp (ada-surgery fork)781031––327682026-10-08community–source
RTX 4070Ternary-Bonsai-2-27B PTQ1_0llama.cpp (ada-surgery fork)69867––655362026-10-08community–source
RTX 4070Ternary-Bonsai-2-27B PTQ1_0llama.cpp (ada-surgery fork)61698––1146882026-10-08community–source
RTX 4070Ternary-Bonsai-2-27B PTQ1_0llama.cpp (ada-surgery fork)30649––1310722026-10-08community–source
RTX 4070Ternary-Bonsai-2-27B PTQ1_0llama.cpp (ada-surgery fork)17.9547––1843202026-10-08community–source
RTX 4070Ternary-Bonsai-2-27B PTQ1_0llama.cpp (ada-surgery fork)8.6436––2621442026-10-08community–source

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

← all hardware