Notes on LLM costs and reliability.
How providers bill, where budgets go wrong, and what we learn building Alye.
· 6 min readWhy your Azure OpenAI and Bedrock bills run 10% above list priceData zones, regional endpoints and cross-region profiles cost more than the global price. What the uplift is, where it applies and how to budget for it.CostAzure OpenAIBedrock · 7 min readPrompt caching: when it saves money and when it costs moreCache writes cost more than plain input and reads cost a tenth. The break-even point, worked examples for Claude and OpenAI, and how to tell which way your traffic goes.CostAnthropicOpenAI · 4 min readIntroducing Alye: know what every LLM call costsWhy we built Alye, what it does, what it deliberately doesn't do, and how to try it on one service in five minutes.Product
Follow along with the RSS feed, or read the longer whitepapers.