AI API cost calculator
Estimate your monthly bill. Enter your typical usage and every model is priced for it, cheapest first.
1,000 tokens ≈ 750 English words. A typical chat message is 50–300 tokens. Token counter →
per 1M tokens · USD. Cached input: the part of each request that repeats (instructions, a document), billed at the cache price where the model has one. Thinking tokens: the hidden reasoning that reasoning models write before answering, billed as output; they only apply to those models.
See how much you save by sending the easy requests to a cheap model.