OpenAI Pricing Calculator
Estimate OpenAI API cost across every GPT model — GPT-5 / 5 Mini / Nano, GPT-4o, GPT-4.1, o3, o3-mini, GPT-3.5. Input + output + cached input pricing, monthly projection at your call volume. Updated against OpenAI's current list prices.
Cost across all GPT models for this workload
How to use the OpenAI Pricing Calculator
Pick a GPT model. Enter your input tokens per call (system prompt + user message), expected output tokens, cache hit ratio (if your system prompt is stable enough to benefit from prompt caching), and monthly call volume. The result panel shows per-call and monthly cost; below, every GPT model is ranked so you can see which is cheapest for your specific workload.
How OpenAI bills
OpenAI bills per million tokens, separately for input and output. Cached input — input tokens that match a prefix you've sent recently — is discounted ~90% on most GPT-5 / GPT-4o models. The catch: caching only kicks in after the first call within a 5-10 minute window, and only for stable prefixes (system message, RAG context that doesn't change per call).
The o-series models bill reasoning tokens as output even though you don't see them. A single o3 call can produce 10-50K reasoning tokens before the visible answer; at $8/M output that adds up fast. Use GPT-5 (which has reasoning built in but bills differently) or GPT-4.1 (no reasoning) if you can't justify it.