If you are interested in Local LLM for Chat, Text generation, but with very simple hardware or you wanna try running LLM on CPU, try https://t.co/VTW7tC6uIR
This model is tuned for Chinese, English, Japanese generation with high quality CPT data and TREMENDOUS SFT data (about 200K lines) from Deepseek V4.1 Flash 1017. Try if you're interested.