AI API Cost Calculators — Tokens to Dollars
The developer side: per-token list prices across the major providers, calculators for per-request and monthly volumes, and a paste-JSON mode that prices any API response in one click.
The price landscape in one paragraph
Per 1M tokens (Oct 2026): the budget tier runs $0.05–0.30 input (GPT-5 nano, Gemini Flash-Lite, DeepSeek V3); the mid tier $0.25–1.25 (GPT-5 mini, GPT-4o mini, GPT-5, Gemini 2.5 Pro); the flagship tier $2–15 input with outputs up to $75 (Claude Opus 4.1). Cached input discounts are the quiet lever — GPT-5 cached input is $0.125/M (−90%), Anthropic cache reads run at 10% of base. The same 10M-in/2M-out workload spans $3.64 (DeepSeek) to $300 (Opus 4.1).
Start here
Have a usage JSON from any provider's API response: paste it into the AI API cost calculator for an instant dollar figure. Budgeting an OpenAI workload: the OpenAI cost calculator prices per-request and batch volumes with the cached-input discount built in. Then read how to reduce OpenAI API costs — the routing lever alone cuts most bills 60%+.