Add throwaway Qwen3.5 q4f16 WebLLM bundle (weights + config + tokenizer + WebGPU .wasm)

initial commit