I made a post about this as well. Most replies were saying this is do to free up hardware resources, but it still doesn’t feel good to lose the option of these models, or having to change a bunch of tools. Specially since OpenAI costs can go up fairly quickly in a few loops with the more robust models.
I’m sorry to say, but according to their update, it seems like we have till January to figure it out and live with the new updates.