What will your AI usage actually cost?
Put in how much you use, see the monthly bill on every major model side by side, and find out whether a $20 subscription or the API is the cheaper way to get it. Prices from the vendors' own pages, checked October 2026. Runs in your browser.
That is 1,200 requests a month: 960K input tokens and 480K output tokens. Nothing here leaves your browser.
| Model | Input $/1M | Output $/1M | Input cost | Output cost | Per month | vs cheapest |
|---|---|---|---|---|---|---|
| GPT-6 LunaOpenAI | $0.10 | $0.50 | $0.10 | $0.24 | $0.34 | cheapest |
| Claude Haiku 5.5Anthropic · * | $0.10 | $0.50 | $0.10 | $0.24 | $0.34 | 1.0× |
| Gemini 3.8 FlashGoogle · * | $0.75 | $3.75 | $0.72 | $1.80 | $2.52 | 7.5× |
| GPT-6.1 SolOpenAI | $2.00 | $10.00 | $1.92 | $4.80 | $6.72 | 20.0× |
| Claude Sonnet 5.5Anthropic | $2.00 | $10.00 | $1.92 | $4.80 | $6.72 | 20.0× |
| Gemini 4 ArgonGoogle · * | $2.00 | $10.00 | $1.92 | $4.80 | $6.72 | 20.0× |
- * Claude Haiku 5.5: Prompts up to 100K tokens; $0.50 / $2.50 above that.
- * Gemini 3.8 Flash: Launch price through 31 December 2026; $1.50 / $7.50 from 1 January 2027.
- * Gemini 4 Argon: Launch price through 30 November 2026; $4 / $20 from 1 December 2026.
Common questions
How is the monthly cost calculated?
Requests per day × 30 days × average tokens per request, split into input and output, multiplied by each model's price per million tokens. Input and output are priced separately because every vendor charges more for output.
What is a token?
Roughly three quarters of an English word. 1,000 tokens is about 750 words. A typical chat question is 20 to 200 tokens; a page of context is about 500; a long document can be tens of thousands.
Where do the prices come from?
From each vendor's own pricing page, checked October 2026: OpenAI, Anthropic and Google. Standard tier, prompts within the normal context window. Batch, cached-input and long-context rates are not included, so real bills can be lower.
Subscription or API, which is cheaper?
The second tab answers that for your usage. A $20 plan is usually cheaper for a person chatting a few dozen times a day; the API is usually cheaper for an app that makes short, predictable calls, and it is the only option for automation.
Does this tool send my numbers anywhere?
No. Everything runs in your browser. Nothing is stored and no AI model is involved.
Sources: OpenAI, Anthropic, Google. Checked October 2026. Prices change; confirm on the vendor's page before you commit.
Find the AI tool for the job.
Most people do not need an API at all. Tell our finder what you want to do and get hand-reviewed tools with real prices and free plans.