Qwen3.6 35B (3B active)

User reports Qwen3.6-35B-A3B at about 90 tokens/sec on an RTX 3070 Laptop 8GB with 32GB DDR4 and an i7-11800H CPU. Setup is PolyStrata with a Q4_K_M GGUF quant; the model is larger than the card's 8GB, so weights ran partly from system RAM. The user describes the low-cost configuration as working quite well for defining tasks, with results that look okay.