Claude API Cost Calculator

This free Claude API cost calculator works out your Claude token cost per request, per day and per month. Pick a Claude model, enter the tokens in each request and how many requests you expect, and it does the math with prompt caching, Batch and Fast mode pricing included.

It also compares every Claude model on the same workload (and OpenAI’s models, if you tick the box) and lists Anthropic’s full API prices, checked against Anthropic’s pricing page on October 18, 2026. Nothing you type leaves your browser.

How to use the Claude API cost calculator

1. Pick a model. Fable 5.1, Opus 5.5, Sonnet 5.5 and Haiku 5.5 are Anthropic’s latest. Older models that are still available are listed below them.
2. Pick Standard, Batch or Fast mode. Batch is half price for jobs that can wait. Fast mode is a premium speed option on Opus 5.5, Opus 5 and Opus 4.8.
3. Enter tokens per request and say where your counts came from. Claude 4.7 and later count about 30% more tokens for the same text, so the calculator adjusts the other models. No counts yet? Paste a sample to estimate.
4. Enter requests per day and days per month to turn the cost per request into a monthly bill.
5. Add prompt caching if part of every prompt repeats, and choose a 5-minute or 1-hour cache.

How Claude API pricing works

Anthropic prices the Claude API per million tokens (MTok), with a different rate for each kind of token:

• Input: everything you send, including the system prompt, chat history and tool definitions.
• Cache writes: storing a prompt prefix costs 1.25 times the input price for a 5-minute cache, or 2 times for a 1-hour cache.
• Cache reads: reusing it costs 0.1 times the input price on most models, 0.05 times on Opus 5.5 and Sonnet 5.5, and 0.025 times on Fable 5.1.
• Output: what Claude writes back, including thinking tokens, at five times the input price on every model in the calculator.

Discounts stack. Batch halves every price, including cache reads and writes, and US-only inference adds 10% on Claude 4.6 and later models.

One more thing changes the bill: Claude 4.7 and later models use a newer tokenizer that produces about 30% more tokens for the same text. When you move from an older model, recount your prompts instead of comparing per-token prices alone.

Claude’s latest models at a glance

Model	Input	Cache read	Output	Batch input / output
Claude Fable 5.1	$10	$0.25	$50	$5 / $25
Claude Opus 5.5	$4	$0.20	$20	$2 / $10
Claude Sonnet 5.5	$2	$0.10	$10	$1 / $5
Claude Haiku 5.5 (prompts up to 100K tokens)	$0.10	$0.01	$0.50	$0.05 / $0.25
Claude Haiku 5.5 (prompts over 100K tokens)	$0.50	$0.05	$2.50	$0.25 / $1.25

Prices per million tokens. The calculator above lists every model, both cache lengths and Fast mode.

Ways to cut your Claude API bill

• Cache your system prompt and documents. A 5-minute cache pays for itself after one read. The repeated part needs a minimum length: 512 tokens on the newest models and up to 4,096 on some older ones.
• Use Batch for anything that can wait. It is half price, most batches finish within an hour, and results are available within 24 hours.
• Try Haiku first. Haiku 5.5 is the cheapest Claude model by far, but its prices rise for prompts over 100,000 tokens.
• Set a max output and a thinking budget. Output, including thinking, costs five times as much as input.
• Recount tokens when you switch models. The newer tokenizer can make a request on a newer model cost more than its per-token price suggests.

Prices exclude tax and negotiated discounts. Claude on Amazon Bedrock and Google Cloud is billed by those platforms at their own prices.

Related on HOW TO // AI

• OpenAI API pricing calculator: the same calculator for OpenAI’s GPT models.
• AI token calculator: paste any text to count its tokens on older and newer Claude models.
• Claude pricing: the Claude app plans, if you want the chat app rather than the API.
• How to pass the Claude Developer exam (CCDV-F): a study plan if you build with the Claude API.

Checked against Anthropic’s pricing, token counting, prompt caching, batch processing, fast mode, extended thinking and model deprecations pages on October 18, 2026.

Frequently asked questions

How much does the Claude API cost per token?

On October 18, 2026, Claude's latest models ranged from $0.10 per million input tokens and $0.50 per million output tokens (Haiku 5.5, for prompts up to 100,000 tokens) to $10 and $50 (Fable 5.1). Sonnet 5.5 costs $2 and $10, and Opus 5.5 costs $4 and $20.

Why do newer Claude models use more tokens?

Anthropic says Claude 4.7 and later models use a newer tokenizer that produces about 30% more tokens for the same text than earlier models. Count your prompts on the model you plan to use before comparing costs.

How much does Claude prompt caching save?

Cache reads cost 10% of the normal input price on most models (5% on Opus 5.5 and Sonnet 5.5, 2.5% on Fable 5.1). Writing to the cache costs 1.25 times the input price for 5 minutes or 2 times for 1 hour, so a 5-minute cache pays off after one read.

How much does the Claude Batch API save?

Batch costs 50% less on input and output for every Claude model. Most batches finish within an hour, and results are available within 24 hours.

Do thinking tokens cost extra on Claude?

Thinking tokens are billed as output tokens, so a large thinking budget raises your output cost. Set a budget and watch your usage to keep it predictable.