For GPT-4.1 mini with OpenAI-only routing, no BYOK, no application retries and a 700-token output limit, what is the maximum billable cost across retries, timeouts and cancellations? Is there a supported hard per-request spending limit?
⌘K
For GPT-4.1 mini with OpenAI-only routing, no BYOK, no application retries and a 700-token output limit, what is the maximum billable cost across retries, timeouts and cancellations? Is there a supported hard per-request spending limit?