GGUF models by win10
97 GGUF models quantized by win10 on Hugging Face. Sizes run from 21M to 40.6B parameters. The longest trained context is 1024K tokens.
About win10 on Hugging Face
win10 has published 97 GGUF models that ggufmodels.com tracks, across families such as Qwen, Phi and Mistral. The most downloaded are DeepSeek-Coder-V2-Lite-Instruct, DeepSeek-Coder-V2-Lite-Instruct and Llama 3.2 3B Instruct 24 9 29.
The quants win10 publishes most often are Q8_0 (64), Q4_K_M (18), Q6_K (11) and Q5_K_M (5). Their Q4_K_M files range from 7.5 to 24.6 GB. Parameter counts run from 21M to 40.6B, and the longest trained context window is 1024K tokens.
Frequently asked questions
What GGUF models has win10 published?
97 models are listed here, mainly from the Qwen, Phi and Mistral families. The most downloaded is DeepSeek-Coder-V2-Lite-Instruct.
Which quants does win10 publish?
The most common are Q8_0 (64 models), Q4_K_M (18 models), Q6_K (11 models) and Q5_K_M (5 models).