Wow, that's a huge increase.

Question: What model are you shifting to, from DeepSeek ? Which is the cheaper model, giving comparable performance to DeepSeek ?

New API pricing just dropped. Output token 2.3, Cache hit flash 2.5, pro: 6.1. Cache miss input 1.5.