Tokenflood allows you to 1) figure out how to slash LLM latency by adjusting prompt parameters 2) assess the load curve of LLM providers before going to production with them