What Is the Anthropic Claude Pricing Calculator?
This calculator helps you estimate the cost of using Anthropic's Claude 3 family of large language models through the Anthropic API. Anthropic is a US-based AI company, and its API prices are published in US dollars (USD). By entering how many tokens your prompt (input) and completion (output) use, and selecting a model, you get an instant cost estimate before you run a job at scale.
Tokens are the small chunks of text that models read and generate. As a rough guide, 1 token is about 4 characters of English, or roughly 0.75 of a word. So 1,000 tokens is approximately 750 words.
How to Use It
- Enter your estimated prompt (input) tokens — the size of your instructions and any context you send.
- Enter your estimated completion (output) tokens — how long you expect Claude's reply to be.
- Choose your model: Claude 3 Opus (most capable), Sonnet (balanced), or Haiku (fastest and cheapest).
- Read the estimated cost for a single request, then multiply by your expected number of requests.
The Formula Explained
Anthropic charges separately for input and output tokens, and output is usually more expensive. The cost is calculated per million tokens:
$$\text{Cost} = \left(\frac{\text{Input tokens}}{1{,}000{,}000}\times \text{Input price}\right) + \left(\frac{\text{Output tokens}}{1{,}000{,}000}\times \text{Output price}\right)$$
Each model has its own input and output rate, so switching from Opus to Haiku can cut costs dramatically for high-volume tasks where top-tier reasoning isn't required.
Worked Example
Suppose you send a 2,000-token prompt and receive a 500-token reply using Claude 3 Sonnet, priced (illustratively) at $3 per million input tokens and $15 per million output tokens:
- Input: \(2{,}000 \div 1{,}000{,}000 \times \$3 = \$0.006\)
- Output: \(500 \div 1{,}000{,}000 \times \$15 = \$0.0075\)
- Total ≈ $0.0135 per request
Run that 10,000 times a month and you'd spend about $135.
Claude 3 Pricing Table (Per Million Tokens)
The Claude 3 model family is priced separately for input (prompt) tokens and output (completion) tokens, billed per million tokens. Opus is the most capable and most expensive; Haiku is the fastest and cheapest. The figures below reflect Anthropic's published list prices for the Claude 3 generation in US dollars.
| Model | Input price (per 1M tokens) | Output price (per 1M tokens) |
|---|---|---|
| Claude 3 Opus | $15.00 | $75.00 |
| Claude 3 Sonnet | $3.00 | $15.00 |
| Claude 3 Haiku | $0.25 | $1.25 |
Note: output tokens cost roughly five times more than input tokens on every Claude 3 model, so completion-heavy workloads (long generated answers) are disproportionately expensive. Prices change over time and may differ by region or volume tier — always verify the current rates on Anthropic's official pricing page before budgeting.
Key Terms Explained
- Token
- The basic unit of text an LLM processes. A token is a chunk of characters — often a word fragment — and for English text roughly 1 token ≈ 4 characters or about 0.75 words. Billing is based on token counts, not characters or words.
- Prompt / input tokens
- All tokens you send to the model in a request, including the system prompt, conversation history, and the current user message. These are billed at the model's input price.
- Completion / output tokens
- The tokens the model generates in its response. These are billed at the (higher) output price, which is typically several times the input price.
- Context window
- The maximum number of tokens — input plus output — the model can consider at once. The Claude 3 family supports a 200,000-token context window. Exceeding it forces you to truncate or summarize earlier content.
- Per-million-token pricing
- The convention Anthropic uses to quote rates: a dollar amount per 1,000,000 tokens. To convert to cost, divide your token count by 1,000,000 and multiply by the per-million price; input and output are priced and totaled separately.
FAQ
Why is output more expensive than input? Generating tokens is more computationally demanding than reading them, so providers price output higher.
Are these the official live prices? Always confirm current rates on Anthropic's pricing page, as prices change over time. Use this tool for planning estimates.
How can I reduce my costs? Trim unnecessary context, cap output length, and use a lighter model like Haiku where high reasoning isn't needed.
