GGUF models by DevQuasar-3
61 GGUF models quantized by DevQuasar-3 on Hugging Face. Sizes run from 14M to 1000B parameters. The longest trained context is 1024K tokens.
About DevQuasar-3 on Hugging Face
DevQuasar-3 has published 61 GGUF models that ggufmodels.com tracks, across families such as Llama, Mistral and DeepSeek. The most downloaded are Llama 3.2 1B, Qwen2.5 Coder 7B Instruct and Deepseek ai.DeepSeek R1 Distill Llama 8B.
The quants DevQuasar-3 publishes most often are Q3_K_M (60), Q2_K (60), Q8_0 (58) and Q5_K_M (58). Their Q4_K_M files range from 0.0 to 604.2 GB. Parameter counts run from 14M to 1000B, and the longest trained context window is 1024K tokens.
Frequently asked questions
What GGUF models has DevQuasar-3 published?
61 models are listed here, mainly from the Llama, Mistral and DeepSeek families. The most downloaded is Llama 3.2 1B.
Which quants does DevQuasar-3 publish?
The most common are Q3_K_M (60 models), Q2_K (60 models), Q8_0 (58 models) and Q5_K_M (58 models).