Is the recently announced conversational memory applied for any model?

Total memory for chats was recently announced. Is this only for ChatGPT 6.1 or will this also work on custom GPTs in the GPT store? So far, we have implemented our own memory store using automation/workflows in n8n.