Drazenko

/

Qwen2.5-Max

like 2

I am going to tell you this: if you want to post something nearly identical to Qwen-2.5-Max, I recommend you study model distillation. I suggest making an account on Alibaba Cloud and getting your Qwen-2.5-Max API. Then you can distill and fine-tune its outputs into this already open-weight model already released. This should be very close performance since it is nearly identical architecture.

Here is the Open-Weight Model I was talking about:

https://huggingface.co/Qwen/Qwen2.5-72B