Phi GGUF models
3486 GGUF models in the Phi family, quantized by 1216 uploaders. Sizes run from 1M to 91.7B parameters. The longest trained context is 128K tokens.
About the Phi GGUF models
Phi is represented on ggufmodels.com by 3,486 GGUF models, quantized by 1,216 uploaders on Hugging Face. The most downloaded are Phi 4 Mini Instruct, Phi 4 Mini Instruct and Phi 3.5 Mini Instruct. Parameter counts run from 1M to 91.7B, and the longest trained context window is 128K tokens.
A Q4_K_M file of a Phi model is between 0.0 to 27.8 GB, which is the quant most people start with. Most files come from mradermacher (660), RichardErkhov (367) and tensorblock (62). The quants you will meet most often are Q4_K_M (2257), Q8_0 (1640), Q5_K_M (1551) and Q6_K (1465), where the number is how many models offer that quant.
Frequently asked questions
Which Phi GGUF should I download?
The most downloaded Phi GGUF here is Phi 4 Mini 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 Phi model need?
Q4_K_M files of Phi models on this site range from 0.0 to 27.8 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 Phi 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.