It's the fate of people relying on cloud services, including the complete removal of old LLM versions.
If you want stability you go local.
Which models do you use locally?
Which models do you use locally?