reddit/r/LocalLLM
Qwen3.8 27b - Holy crap!!
The user wants a concise HTML summary of the Reddit post about Qwen3.8 27b. The post describes the user's experience running Qwen3.8 27b IQ4 NL at medium reasoning effort on their hardware (4060 8GB x4, 5060 Ti 16GB x8, 24GB VRAM total, i9-10900, 32GB RAM). They achieve ~20 tok/s at 128k context using llama.cpp backend for OpenWebUI with OpenTerminal MCP. They asked Qwen to create Pong using Python, but the description cuts off. I need to summarize in 2-4 sentences, precise technical language, only from provided info. Output only HTML with