While Chinese models train on all Western cultural output, our own models are restricted. And in the corporate world the models of choice for finetuning are DeepSeek and Qwen, wonder why.