Originally published on Kael Research

TL;DR: Below 1B tokens/month, just use APIs. Proprietary or hosted open-source, doesn't matter much. Between 1 and 10B tokens, hosted open-source APIs from Together.ai or Groq are usually cheapest. Above 10B tokens/month, self-hosting can win, but only if you already have an MLOps team. The "open source is...