Llamacpp imatrix Quantizations of Qwen3 4B by Qwen Using llama.cpp release b5200 for quantization. Original model: https://huggingface.co/Qwen/Qwen3 4B All quants made using imatrix option with dataset from here Run them in LM Studio Run them directly with llama.cpp, or any other llama.cpp based project Prompt format Download a file (not the whole branch) from below: Filename Quant type File Size Split Description Qwen3 4B bf16.gguf bf16 8.05GB false Full BF16 weights. Qwen3 4B Q8 0.gguf Q8 0 4.28GB false Extremely high quality, generally unneeded but max available quant. Qwen3 4B Q6 K L.gguf Q6 K L 3.40GB false Uses Q8 0 for embed and output weights. Very high quality, near perfect, recommended . Qwen3 4B Q6 K.gguf Q6 K 3.31GB false Very high quality, near perfect, recommended . Qwen3 4B Q5 K L.gguf Q5 K L 2.98GB false Uses Q8 0 for embed and output weights. High quality, recommended . Qwen3 4B Q5 K M.gguf Q5 K M 2.89GB false High quality, recommended . Qwen3 4B Q5 K S.gguf Q5 K S 2.82GB false High quality, recommended . Qwen3 4B Q4 1.gguf Q4 1 2.60GB false Legacy format, similar performance to Q4 K S but with improved tokens/watt on Apple silicon. Qwen3 4B Q4 K L.gguf Q4 K L 2.59…
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy