Your statement comes in extreme contrast with my experience using local LLMs for more than year now. Qwen 35B-A3B, even its Q4 quantization, is extremely capable. Hell, at this point i pretty much turn on my PC, then run llama-cpp just to have it in the background for when i need it to do stuff.

My local model (Qwen 3.6 27b) with a basic web search tool call has proven a fantastic “dumb” assistant editor/writer when writing. It’s really breaks through my writers block.

It’s limited, it’s a bit finicky here and there, but it’s been very helpful. Mostly just gives me nuggets of ideas or shows me what I don’t want, which is informative.

i really like the idea of running local models but i'm always in the position of wanting the best model(s) available and i don't have any severe privacy concerns. as such i have yet to justify ever using local models.