The Qwen2.5 72B model is constantly overloaded. We need a new, fast, and efficient model to be the default!
Mistral small 3.1, perhaps? Wait can't we use Gemma3?
· Sign up or log in to comment
