LLM API cost calculator
What an AI feature costs a month on current OpenAI, Claude and Gemini models. Set how many requests you expect and how many tokens go in and come out of each; every price can be edited.
| Model | $ / 1M input | $ / 1M output | Per request | Per month |
|---|---|---|---|---|
| GPT-6 LunaOpenAI · Prompts up to 272K tokens | $0.0004 | $40 | ||
| Claude Haiku 5.5Anthropic · Prompts up to 100K tokens | $0.0004 | $40 | ||
| Gemini 3.5 Flash-LiteGoogle · Text, image and video input | $0.0016 | $160 | ||
| Gemini 3.8 FlashGoogle · Until 31 Dec 2026; $1.50 / $7.50 from 1 Jan 2027 | $0.003 | $300 | ||
| GPT-6.1 SolOpenAI · Prompts up to 272K tokens | $0.008 | $800 | ||
| Claude Sonnet 5.5Anthropic · Full 1M context at this price | $0.008 | $800 | ||
| Gemini 3.1 Pro (preview)Google · Prompts up to 200K tokens | $0.0088 | $880 | ||
| Claude Opus 5.5Anthropic · Full 1M context at this price | $0.02 | $1,600 | ||
| GPT-6 AstraOpenAI · Prompts up to 272K tokens | $0.04 | $4,000 | ||
| Claude Fable 5.1Anthropic · Full 1M context at this price | $0.04 | $4,000 |
Standard prices as of 8 October 2026, from OpenAI, Anthropic and Google. They change, so check before you budget.
How the estimate works
Each request costs its input tokens times the input price, plus its output tokens times the output price, with prices quoted per million tokens. The monthly figure is that, times the number of requests. Input is everything sent to the model: the system prompt, conversation history, any retrieved documents and the question itself. As a rough guide, a token is about three-quarters of an English word.
What it leaves out
- Prompt caching and batch discounts, which can cut the bill well below these figures.
- Higher prices some models charge for very long prompts (noted beside each model).
- Retries, tool calls and agent loops that call the model several times per task.
- Embeddings, vector storage and hosting.
For what drives the number up in a real product, and how to bring it down, read what an AI feature costs to run. If your bill is already growing faster than your usage, an LLM cost and RAG quality audit measures it on your real traffic.
Tell me what you’re building and where it’s stuck.
I’ll tell you the cleanest path forward, including if it’s “don’t build that.”
Or write tocontact@alihassan.dev
