RTX 4070 Ti SUPER

20 records across 6 models (quants: EXL2 8.0bpw, MXFP4, PTQ1_0, Q4_0, Q4_K, Q8_0).

hardwaremodelquantbackendtok/spp tok/sWttft sctx/tgdateclassflagssource
RTX 4070 Ti SUPERgpt-oss 20B 20BMXFP4llama.cpp128.95628.1––4096–sourced–source
RTX 4070 Ti SUPERgpt-oss 20B 20BMXFP4llama.cpp113.64182.7––16384–sourced–source
RTX 4070 Ti SUPERgpt-oss 20B 20BMXFP4llama.cpp98.73020.0––32768–sourced–source
RTX 4070 Ti SUPERgpt-oss 20B 20BMXFP4llama.cpp78.71716.2––65536–sourced–source
RTX 4070 Ti SUPERgpt-oss 20B 20BMXFP4llama.cpp57.5993.5––131072–sourced–source
RTX 4070 Ti SUPERLlama 3 8B Instruct 8BEXL2 8.0bpwExLlamaV264–––81922024-05community–source
RTX 4070 Ti SUPERLlama 3 8B Instruct 8BQ8_0LM Studio56–––81922024-05community–source
RTX 4070 Ti SUPERllama 7B Q4_0llama.cpp (Vulkan)135.816801.18––tg1282025-09-24community–source
RTX 4070 Ti SUPERllama 7B Q4_0llama.cpp (Vulkan)129.456099.18––tg1282025-09-24community–source
RTX 4070 Ti SUPERQwen3 14B 14BQ4_Kllama.cpp58.13048.0––4096–sourced–source
RTX 4070 Ti SUPERQwen3 14B 14BQ4_Kllama.cpp47.22003.4––16384–sourced–source
RTX 4070 Ti SUPERQwen3 14B 14BQ4_Kllama.cpp37.71236.9––32768–sourced–source
RTX 4070 Ti SUPERQwen3 8B 8BQ4_Kllama.cpp96.35220.1––4096–sourced–source
RTX 4070 Ti SUPERQwen3 8B 8BQ4_Kllama.cpp72.23050.7––16384–sourced–source
RTX 4070 Ti SUPERQwen3 8B 8BQ4_Kllama.cpp54.51616.6––32768–sourced–source
RTX 4070 Ti SUPERQwen3 8B 8BQ4_Kllama.cpp36.6829.4––65536–sourced–source
RTX 4070 Ti SUPERTernary-Bonsai-2-27B PTQ1_0llama.cpp (xyz-llama fork)156.3–––1638402026-09community–source
RTX 4070 Ti SUPERTernary-Bonsai-2-27B PTQ1_0llama.cpp (xyz-llama fork)144.2–––1638402026-09community–source
RTX 4070 Ti SUPERTernary-Bonsai-2-27B PTQ1_0llama.cpp (xyz-llama fork)141.3–––1638402026-09community–source
RTX 4070 Ti SUPERTernary-Bonsai-2-27B PTQ1_0llama.cpp (xyz-llama fork)139.5–––1638402026-09community–source

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

← all hardware