Our LLM cost model was double-counting batch and cache discounts(beforeyouship.dev)
1 pointby indiegoing9 minutes ago1 comment