[Feature Request] Consistent model performance and transparency in capability changes

Question

**Bug Description**
As a user it's very obvious when you are nerfing the model. The moment Fable came out, Opus became dumb. Then Fable got suspended, Opus became smart again. And I'm not sure what the internal plan is now? Maybe trying to push us to /fast, but opus is so dumb again. It's so annoying. How is that a reasonable strategy. Put the user first. I understand it's maybe not sustainable, but this kind of low quality strategy is going to backfire so badly.

**Environment Info**
- Platform