Remember nodes and graphs? A comfy user interface allows pretty incredible wiring among models local ai is like eurorack. The current graph skews heavily towards a a pair of small dense models collaborating with the large heavyweights selectively. It’s Qwen 3.6 27B with Gemma 4 31B, both unquantized, bf16/fp16, with phi 14b, nemotron cascade 2, and then those large heavyweights, r1 and subsequent deepseek models including speciale, gpt oss 120b, glm, min max,kimi, command r, mistrals, ever body, up in one graph, all them llm nodes patched and interconnected. Slow, resource intense, better than non local ai. I used Matteo’s graphllm for inspiration, and comfy ui (and st), and used the models to roll a new imgui node/graph model compositor. Now what?!