What does your prompt actually cost?

Paste the prompt you send on every call. See the monthly bill on 14 models side by side — exact token counts for OpenAI, hand-checked prices for everyone.

Some prices below have not been hand-verified yet. Treat them as indicative until the checkmark appears.

ModelInput / 1MOutput / 1MCost / month
GPT-4.1 nanoOpenAI$0.100 / 1M$0.400 / 1M$2.05
Gemini 2.5 Flash-LiteGoogle$0.100 / 1M$0.400 / 1M$2.06
Gemini 2.0 FlashGoogle$0.100 / 1M$0.400 / 1M$2.06
GPT-4o miniOpenAI$0.150 / 1M$0.600 / 1M$3.07
GPT-4.1 miniOpenAI$0.400 / 1M$1.60 / 1M$8.18
Gemini 2.5 FlashGoogle$0.300 / 1M$2.50 / 1M$12.68
Claude Haiku 3.5Anthropic$0.800 / 1M$4.00 / 1M$20.48
o4-miniOpenAI$1.10 / 1M$4.40 / 1M$22.51
Claude Haiku 4.5Anthropic$1.00 / 1M$5.00 / 1M$25.60
GPT-4.1OpenAI$2.00 / 1M$8.00 / 1M$40.92
Gemini 2.5 ProGoogle$1.25 / 1M$10.00 / 1M$50.75
GPT-4oOpenAI$2.50 / 1M$10.00 / 1M$51.15
Claude Sonnet 4.5Anthropic$3.00 / 1M$15.00 / 1M$76.80
Claude Opus 4.1Anthropic$15.00 / 1M$75.00 / 1M$384

Figures assume standard synchronous API pricing. Prompt caching, batch discounts, and reasoning-token overhead are not modelled. Output-token count is your estimate, not a measurement.

Where the numbers come from

Prices are transcribed by hand from each provider’s public pricing page and re-checked regularly. If you spot a stale number, it is a bug — tell me.