CheckPre

Claude Haiku 4.5 API Cost Calculator

What Claude Haiku 4.5 costs at $1 in and $5 out per million tokens, for high-volume work where speed and price matter more than depth.

Updated · By Niraj Kumar · 6 sources

Each request

Count text inCount text in
tokens

System prompt, conversation history, documents and tool definitions are all resent on every call.

tokens

The visible answer.

Count tokens from a prompt or JSON payloadShow

Plain text, OpenAI Chat or Responses JSON, or Anthropic Messages JSON. A final assistant message or a response body counts as output. Counted in your browser; nothing is uploaded.

Loading the tokenizer…

Counts use OpenAI's o200k tokenizer, so they are exact for GPT-6 and GPT-5 models, plus a few tokens of message framing. Other providers' tokenizers count the same text differently and do not publish a conversion factor; Anthropic says Claude models from 4.7 on produce about 30% more tokens than earlier Claude models. For an exact count on another model, paste one of its API responses: the usage it reports is shown above.

tokens

Thinking tokens are billed as output. Leave at 0 for models or settings that do not reason.

Volume

PricePrice

30,000 requests over 30 days.

Discounts

%
0%100%

The share of input that repeats exactly at the start of every request, such as a fixed system prompt. The prompt caching calculator prices cache writes and expiry in detail.

Batch APIBatch API
Monthly cost$165.00$0.0055 a request on Claude Haiku 4.5
Per 1,000 requests
$5.50
Input
$90.00
Output
$75.00
Price rank
10th cheapest of 24
Cheapest listed model
Claude Haiku 5.5
$16.50 a month, $148.50 less
Output45%
  • Input55%
  • Output45%
Every model on this workload, per month3,000 tokens in, 500 out, 1,000 a day. Hover or focus a row for rates.
  1. 1Claude Haiku 5.5Anthropic$16.50
  2. 2GPT-6 LunaOpenAI$16.50
  3. 3Qwen3.8 FlashAlibaba$20.55
  4. 4GPT-5.6 LunaOpenAI$36.00
  5. 5DeepSeek V4.1 FlashDeepSeek$45.00
  6. 6Qwen3.7 PlusAlibaba$60.00
  7. 7Gemini 3.5 Flash-LiteGoogle$64.50
  8. 8Gemini 3.8 FlashGoogle$123.75
  9. 9Grok 4.3xAI$150.00
  10. 10Claude Haiku 4.5Anthropic$165.00
  11. 11DeepSeek V4 ProDeepSeek$178.20
  12. 12Grok 4.7xAI$270.00
  13. 13Qwen3.8 MaxAlibaba$270.00
  14. 14Claude Sonnet 5.5Anthropic$330.00
  15. 15Claude Sonnet 5Anthropic$330.00
  16. 16GPT-6.1 SolOpenAI$330.00
  17. 17GPT-6 SolOpenAI$330.00
  18. 18GPT-5.6 TerraOpenAI$360.00
  19. 19Gemini 3.1 ProGoogle$360.00
  20. 20Claude Opus 5.5Anthropic$660.00
  21. 21GPT-5.6 SolOpenAI$660.00
  22. 22Claude Opus 5Anthropic$825.00
  23. 23Claude Fable 5.1Anthropic$1,650
  24. 24GPT-6 AstraOpenAI$1,650

How the Claude Haiku 4.5 API Cost Calculator works

As of October 13, 2026, Claude Haiku 4.5 costs $1 per million input tokens and $5 per million output tokens, with cache hits at $0.10. A request costs its input tokens × the input rate ÷ 1,000,000 plus its output tokens × the output rate ÷ 1,000,000; multiply by requests for the monthly bill.

cost = (fresh input × input rate + cached input × cache rate + (output + reasoning) × output rate) ÷ 1,000,000

This page uses the same method as the LLM API Cost Calculator, which explains each term and assumption in full.

Worked example

A chat feature on Claude Haiku 4.5 sending 3,000 input tokens and getting 500 back, 1,000 times a day.

  1. 1Input: 3,000 × $1 ÷ 1,000,000 = $0.003
  2. 2Output: 500 × $5 ÷ 1,000,000 = $0.0025
  3. 3Per request: $0.0055
  4. 4Per month: $0.0055 × 1,000 × 30 = $165.00

$165.00 a month. Caching a fixed system prompt that makes up 80% of the input would bring it to about $100.20, and the batch API would halve either figure.

Frequently asked questions

How much does Claude Haiku 4.5 cost per million tokens?

$1 per million input tokens and $5 per million output tokens at Anthropic's standard rate. Cache hits are $0.10, and writing to the cache costs $1.25 for five minutes or $2 for an hour. The batch API takes 50% off. One rate covers its full 200,000-token context window.

What limits should I know about on Claude Haiku 4.5?

Its context window is 200,000 tokens and it writes at most 64,000 tokens per response, smaller than the 1M context of newer Claude models; the calculator flags requests that will not fit. Its minimum cacheable prompt is 4,096 tokens, so short prompts will not cache. It predates the tokenizer Claude 4.7 introduced, so the same text usually counts as fewer tokens than on newer Claude models.

How big a request can Claude Haiku 4.5 take?

Up to 200,000 tokens of context, input and output together, and up to 64,000 output tokens per response. That is roughly 150,000 words of English. The calculator warns when a request will not fit.

Sources

Related calculators

More in AI, Cloud & Developer