Hi hunters! I built this because every "how much does the OpenAI API cost" article I found was either outdated or did naive math (price × requests) that misses how APIs actually bill. What it does differently: - Prices for 292 models refresh automatically every day, so a model that launched this morning has a page today. - The calculators simulate real billing mechanics: chatbots resend growing history every turn, agents resend tool schemas every step, caching discounts the repeated prefix by ~90%. The agent calculator shows how a task that "feels like" one request is often 15 requests — real bills come out 10–30× above naive estimates. - There's also a GPU rental comparison (same H100 spans $1.49–$6.98/hr across providers), 9 cost guides, and a token counter that runs entirely client-side. It's a static site (Astro), free, no signup. Would love feedback on what cost questions you'd want answered that aren't covered yet.

CostPerPrompt

Live AI API pricing + calculators that model real workloads

Live AI API pricing + calculators that model real workloads