Q3_K_M GGUF models
3-bit k-quant, medium. Small files with some quality loss. 108680 models have a Q3_K_M file. Sizes run from 1M to 2780B parameters. The longest trained context is 4194304K tokens.
What is Q3_K_M?
3-bit k-quant, medium. Small files with some quality loss. On average it stores about 3.9 bits per weight, so a Q3_K_M file is roughly 24% of the size of the same model in 16-bit. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.
108,680 models have a Q3_K_M file on ggufmodels.com, and those files range from 0.0 to 1325.0 GB. The most downloaded are Ornith 1.5 9B Uncensored, Qwen Image 2.1 and Gemma-4-12B-It. For a longer explanation of every quant type, read GGUF quantization explained.
Frequently asked questions
What is Q3_K_M in GGUF?
3-bit k-quant, medium. Small files with some quality loss. It averages about 3.9 bits per weight.
Should I use Q3_K_M?
A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.
How big is a Q3_K_M file?
Q3_K_M files on this site range from 0.0 to 1325.0 GB, depending on the model. As a rule of thumb it is about 24% of the 16-bit size.