This looks great, but I don't see any mention on the performance. TPM rate limit is almost x3, which helps a lot of non-conversational tasks, so it would be a h

This looks great, but I don't see any mention on the performance. TPM rate limit is almost x3, which helps a lot of non-conversational tasks, so it would be a huge upgrade if the output quality can be equivalent or better than the chat model.