3 ms·
I'm running qwen36.:35b:iq4 IQ4_XS quant. Takes 18 GB of RAM with 131k context window. Seems to be really good. Have it running local stuff via Hermes, using
by diseasedyak 3mo ago
I'm running qwen36.:35b:iq4 IQ4_XS quant. Takes 18 GB of RAM with 131k context window. Seems to be really good. Have it running local stuff via Hermes, using a cloud model via Ollama (Deepseek V4-Pro) for heavy lifting.