Tesla P40

41 records across 3 models (quants: EXL2 4.0bpw H6, Q2_K, Q3_K_M, Q4_0, Q4_K_M).

hardwaremodelquantbackendtok/spp tok/sWttft sctx/tgdateclassflagssource
Tesla P40CodeLlama-34B 34BEXL2 4.0bpw H6ExLlamaV21.1905–80––2023-09community–source
Tesla P40Llama 3.1 8B Q4_0llama.cpp (CUDA)54.75–––tg1282026-10community–source
Tesla P40Llama 3.1 8B Q4_0llama.cpp (CUDA)53.69–––4962026-10community–source
Tesla P40Llama 3.1 8B Q4_0llama.cpp (CUDA)52.28–––tg1282026-10community–source
Tesla P40Llama 3.1 8B Q4_0llama.cpp (CUDA)51.87–––20162026-10community–source
Tesla P40Llama 3.1 8B Q4_0llama.cpp (CUDA)51.39–––4962026-10community–source
Tesla P40Llama 3.1 8B Q4_0llama.cpp (CUDA)49.72–––20162026-10community–source
Tesla P40Llama 3.1 8B Q4_K_Mllama.cpp (CUDA)48.94–––tg1282026-10community–source
Tesla P40Llama 3.1 8B Q4_K_Mllama.cpp (CUDA)48.34–––tg1282026-10community–source
Tesla P40Llama 3.1 8B Q4_K_Mllama.cpp (CUDA)48.04–––4962026-10community–source
Tesla P40Llama 3.1 8B Q4_K_Mllama.cpp (CUDA)47.47–––4962026-10community–source
Tesla P40Llama 3.1 8B Q4_K_Mllama.cpp (CUDA)46.53–––20162026-10community–source
Tesla P40Llama 3.1 8B Q4_K_Mllama.cpp (CUDA)46.05–––20162026-10community–source
Tesla P40Mixtral 8x7B Q2_Kllama.cpp (CUDA)32.42–––tg1282026-10community–source
Tesla P40Mixtral 8x7B Q2_Kllama.cpp (CUDA)32.03–––4962026-10community–source
Tesla P40Mixtral 8x7B Q2_Kllama.cpp (CUDA)31.37–––20162026-10community–source
Tesla P40Mixtral 8x7B Q2_Kllama.cpp (CUDA)30.65–––tg1282026-10community–source
Tesla P40Mixtral 8x7B Q2_Kllama.cpp (CUDA)30.30–––4962026-10community–source
Tesla P40Mixtral 8x7B Q2_Kllama.cpp (CUDA)29.69–––20162026-10community–source
Tesla P40Mixtral 8x7B Q3_K_Mllama.cpp (CUDA)27.12–––tg1282026-10community–source
Tesla P40Mixtral 8x7B Q3_K_Mllama.cpp (CUDA)26.83–––4962026-10community–source
Tesla P40Mixtral 8x7B Q3_K_Mllama.cpp (CUDA)26.35–––20162026-10community–source
Tesla P40Mixtral 8x7B Q3_K_Mllama.cpp (CUDA)25.76–––tg1282026-10community–source
Tesla P40Mixtral 8x7B Q3_K_Mllama.cpp (CUDA)25.54–––4962026-10community–source
Tesla P40Mixtral 8x7B Q3_K_Mllama.cpp (CUDA)25.06–––20162026-10community–source
Tesla P40Mixtral 8x7B Q2_Kllama.cpp (CUDA)–33.71–––2026-10community–source
Tesla P40Mixtral 8x7B Q2_Kllama.cpp (CUDA)–31.82–––2026-10community–source
Tesla P40Mixtral 8x7B Q2_Kllama.cpp (CUDA)–46.65–––2026-10community–source
Tesla P40Mixtral 8x7B Q2_Kllama.cpp (CUDA)–45.35–––2026-10community–source
Tesla P40Mixtral 8x7B Q2_Kllama.cpp (CUDA)–51.35–––2026-10community–source
Tesla P40Mixtral 8x7B Q2_Kllama.cpp (CUDA)–50.64–––2026-10community–source
Tesla P40Mixtral 8x7B Q2_Kllama.cpp (CUDA)–54.41–––2026-10community–source
Tesla P40Mixtral 8x7B Q2_Kllama.cpp (CUDA)–54.33–––2026-10community–source
Tesla P40Mixtral 8x7B Q3_K_Mllama.cpp (CUDA)–26.74–––2026-10community–source
Tesla P40Mixtral 8x7B Q3_K_Mllama.cpp (CUDA)–28.14–––2026-10community–source
Tesla P40Mixtral 8x7B Q3_K_Mllama.cpp (CUDA)–32.71–––2026-10community–source
Tesla P40Mixtral 8x7B Q3_K_Mllama.cpp (CUDA)–36.54–––2026-10community–source
Tesla P40Mixtral 8x7B Q3_K_Mllama.cpp (CUDA)–35.99–––2026-10community–source
Tesla P40Mixtral 8x7B Q3_K_Mllama.cpp (CUDA)–40.58–––2026-10community–source
Tesla P40Mixtral 8x7B Q3_K_Mllama.cpp (CUDA)–57.55–––2026-10community–source
Tesla P40Mixtral 8x7B Q3_K_Mllama.cpp (CUDA)–57.53–––2026-10community–source

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

← all hardware