Your session has been flagged for unusual activity.
Please confirm you're human to continue.
This LLM cost calculator shows what a prompt costs on the API for GPT, Claude, Gemini and Grok, per call and per month. Enter your input and output tokens for an instant LLM API cost comparison at current list prices.
Not sure how many tokens your prompt is? Paste it into the token counter. About 750 English words ≈ 1,000 tokens.
| Model$/1M in / out | $/1M in | $/1M out | Per call | Monthly |
|---|
Standard pay-as-you-go list prices in USD, sorted cheapest first. Caching and batch discounts aren't included.
Every major provider bills the API by the token, with two separate prices quoted per million tokens:
Input tokens
Everything you send: system prompt, chat history, documents and your question.
Output tokens
Everything the model writes back. Usually 5× the input price; 2–3× on Grok.
Per 1M tokens
A million tokens is roughly 750,000 English words, so single calls cost fractions of a cent.
The formula: cost per call = (input tokens ÷ 1,000,000 × input price) + (output tokens ÷ 1,000,000 × output price). Multiply by calls per month for your monthly bill.
What a simple estimate leaves out:
A chatbot sends 2,000 input tokens (instructions, history and the customer's question) and gets 500 output tokens back, 10,000 times a month, on Claude Opus 5.5 at $4 input and $20 output per 1M tokens.
The same workload at October 2026 prices on other models:
$4.50
GPT-6 Luna / month
$70
Grok 4.7 / month
$90
Claude Sonnet 5.5 / month
$450
Claude Fable 5.1 / month
ChatGPT Plus and Claude Pro cost $20 a month and Google AI Pro $19.99. Those plans are for using the chat apps; API usage is billed separately, per token. Which is cheaper depends on how you work.
A subscription usually wins when
The API usually wins when
The break-even math: $20 of API usage buys about 571 GPT-6 Astra messages of 1,000 input and 500 output tokens, or about 2,857 on Claude Sonnet 5.5. Once each message carries 20,000 tokens of chat history and files, the same $20 covers only about 89 GPT-6 Astra messages. Subscriptions aren't unlimited either: ChatGPT Plus is metered over five-hour and weekly windows and Claude Pro over a rolling five-hour session (see limits by plan).
Standard list prices per 1 million tokens, as used by the calculator above.
| Model | Input / 1M | Output / 1M |
|---|---|---|
| GPT-6 AstraOpenAI | $10.00 | $50.00 |
| GPT-6.1 SolOpenAI | $2.00 | $10.00 |
| GPT-6 LunaOpenAI | $0.10 | $0.50 |
| Claude Fable 5.1Anthropic | $10.00 | $50.00 |
| Claude Opus 5.5Anthropic | $4.00 | $20.00 |
| Claude Sonnet 5.5Anthropic | $2.00 | $10.00 |
| Claude Haiku 5.5Anthropic | $0.10 | $0.50 |
| Gemini 4 ArgonGoogle · $4.00 / $20.00 from Dec 1, 2026 | $2.00 | $10.00 |
| Gemini 3.8 FlashGoogle · $1.50 / $7.50 from Jan 1, 2027 | $0.75 | $3.75 |
| Grok 4.7xAI | $2.00 | $6.00 |
| Grok 4.3xAI | $1.25 | $2.50 |
USD, standard tier. Sources: OpenAI API pricing, Anthropic pricing (platform.claude.com), Gemini API pricing and xAI docs, checked October 2026.
Multiply your input tokens by the input price and your output tokens by the output price, both per million tokens, and add the two. For example, 2,000 input and 500 output tokens on Claude Opus 5.5 ($4 and $20 per 1M) cost 0.002 x $4 + 0.0005 x $20 = $0.018 per call. Multiply by your monthly calls for the monthly bill.
At October 2026 list prices, GPT-6 Luna and Claude Haiku 5.5 are the lowest in this calculator at $0.10 per 1M input and $0.50 per 1M output tokens. Among the larger models, Grok 4.3 ($1.25 and $2.50) and Gemini 3.8 Flash ($0.75 and $3.75 until December 31, 2026) are the cheapest. The cheapest model is not always the best value if it needs retries or extra clean-up.
The model reads your input in one pass but generates the answer one token at a time, which takes more compute. For most models here output costs 5 times the input price; for Grok 4.7 it is 3 times and for Grok 4.3 2 times. On reasoning models the thinking tokens are usually billed as output too.
In English, one token is about 4 characters or three-quarters of a word, so a 750-word prompt is roughly 1,000 tokens. For an exact GPT count, paste the text into the token counter, which runs OpenAI's tokenizer in your browser.
It depends on how much you use it. $20 of API usage buys about 571 GPT-6 Astra messages of 1,000 input and 500 output tokens, or about 89 once each message carries 20,000 tokens of chat history and files. Heavy daily chat with top models is usually cheaper on a $20 subscription, while light or automated use with smaller models is usually cheaper on the API. Subscriptions have their own usage limits and don't include API access.
Google lists Gemini 4 Argon at $2 input and $10 output per 1M tokens through November 30, 2026, and $4 and $20 from December 1, 2026. Gemini 3.8 Flash is $0.75 and $3.75 through December 31, 2026, and $1.50 and $7.50 from January 1, 2027. This calculator switches to the new prices automatically on those dates.
No. It uses each provider's standard pay-as-you-go rates. Most providers offer lower prices for cached input and for batch jobs, and charge separately for images, audio, web search and other tools, so your real bill can be lower or higher than the estimate.
They are public list prices checked in October 2026 on the OpenAI, Anthropic, Google and xAI pricing pages. Prices change often, so confirm on the provider's pricing page before committing to a budget.
MultipleChat Pro is $20 a month for ChatGPT, Claude, Gemini, Grok and Perplexity in one workspace, with side-by-side answers and no API keys to manage.
Try MultipleChat