Add Qwen models for text generation

I have been using Qwen 2.5-72B on another platform, and I’d love to see it on Venice. You already have a Qwen model, but it is focused on coding. I find Qwen to be more natural and precise when summarising long texts or when asked to write an essay compared to Llama and DeepSeek.