Loading request...
[Qwen] Allow switching between thinking and coding modes without reloading the model | RequestHunt