Llamacpp imatrix Quantizations of Mistral Medium 3.5 128B by mistralai Using llama.cpp release b9010 for quantization. Original model: https://huggingface.co/mistralai/Mistral Medium 3.5 128B All quants made using imatrix option with dataset from here Run them in your choice of tools: llama.cpp ramalama LM Studio koboldcpp Jan AI Text Generation Web UI LoLLMs Note: if it's a newly supported model, you may need to wait for an update from the developers. Prompt format What's new: Fix for rope settings (needed new imatrix as well) and fixed mmproj (currently only bf16 and f32 seem to work, f16 gives incoherent results) Download a file (not the whole branch) from below: Filename Quant type File Size Split Description Mistral Medium 3.5 128B Q8 0.gguf Q8 0 132.85GB true Extremely high quality, generally unneeded but max available quant. Mistral Medium 3.5 128B Q6 K L.gguf Q6 K L 108.58GB true Uses Q8 0 for embed and output weights. Very high quality, near perfect, recommended . Mistral Medium 3.5 128B Q6 K.gguf Q6 K 107.80GB true Very high quality, near perfect, recommended . Mistral Medium 3.5 128B Q5 K L.gguf Q5 K L 92.11GB true Uses Q8 0 for embed and output weights. High quality, re…
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy