Llama GGUF models
30022 GGUF models in the Llama family, quantized by 7347 uploaders. Sizes run from 1M to 1020B parameters. The longest trained context is 102400K tokens.
About the Llama GGUF models
Llama is represented on ggufmodels.com by 30,022 GGUF models, quantized by 7,347 uploaders on Hugging Face. The most downloaded are Meta Llama 3.1 8B Instruct, Llama 3.2 3B Instruct and Llama-3.2-3B-Instruct. Parameter counts run from 1M to 1020B, and the longest trained context window is 102400K tokens.
A Q4_K_M file of a Llama model is between 0.0 to 245.7 GB, which is the quant most people start with. Most files come from mradermacher (8238), RichardErkhov (3423) and tensorblock (1093). The quants you will meet most often are Q4_K_M (21448), Q5_K_M (16056), Q2_K (15979) and Q8_0 (15312), where the number is how many models offer that quant.
Frequently asked questions
Which Llama GGUF should I download?
The most downloaded Llama GGUF here is Meta Llama 3.1 8B Instruct. Pick the size that fits your GPU first, then choose a quant: Q4_K_M is the usual starting point, and Q5_K_M or Q6_K give more quality if you have memory to spare. The VRAM pages below list what fits each GPU.
How much VRAM does a Llama model need?
Q4_K_M files of Llama models on this site range from 0.0 to 245.7 GB. Add the KV cache for your context length and about 0.5 GB for the runtime, which usually means 1.5 to 3 GB on top of the file at an 8K context.
Where do Llama GGUF files come from?
The most common quantizers here are mradermacher, RichardErkhov and tensorblock. They convert the original model and publish quantized files on Hugging Face. ggufmodels.com does not host files; every download button opens Hugging Face.