Developer · Anthropic

Claude Haiku 4.5 Token Counter & Cost Calculator

Use this Claude Haiku 4.5 token counter to see how many tokens a prompt uses and what it costs: $1.00 per million input tokens and $5.00 per million output tokens. Paste text below — it is counted in your browser and never uploaded.

0 tokens0 characters0 words
ModelInput tokensInput costOutput costPer callPer monthContext
Claude Haiku 4.5Anthropic≈0$0$0.0025$0.0025$75.00200k
GPT-5 nanoOpenAIcheapest≈0$0$0.0002$0.0002$6.00400k
Gemini 2.5 Flash-LiteGooglecheapest≈0$0$0.0002$0.0002$6.001M
GPT-6 LunaOpenAI≈0$0$0.00025$0.00025$7.50—
GPT-4o miniOpenAI0$0$0.0003$0.0003$9.00128k
GPT-5.6 LunaOpenAI≈0$0$0.0006$0.0006$18.00—
GPT-5 miniOpenAI≈0$0$0.001$0.001$30.00400k
Gemini 3.5 Flash-LiteGoogle≈0$0$0.00125$0.00125$37.501M
Gemini 2.5 FlashGoogle≈0$0$0.00125$0.00125$37.501M
Gemini 3.8 FlashGoogle≈0$0$0.001875$0.001875$56.251M
GPT-5.4 miniOpenAI≈0$0$0.00225$0.00225$67.50—
GPT-4.1OpenAI0$0$0.004$0.004$1201.05M
Gemini 3.5 FlashGoogle≈0$0$0.0045$0.0045$1351M
Claude Sonnet 5Anthropic≈0$0$0.005$0.005$1501M
GPT-6 SolOpenAI≈0$0$0.005$0.005$150—
GPT-4oOpenAI0$0$0.005$0.005$150128k
Gemini 2.5 ProGoogle≈0$0$0.005$0.005$1501M
GPT-5.6 TerraOpenAI≈0$0$0.006$0.006$180—
Gemini 3.1 Pro PreviewGoogle≈0$0$0.006$0.006$180200k
Claude Sonnet 4.6Anthropic≈0$0$0.0075$0.0075$2251M
Claude Opus 5.5Anthropic≈0$0$0.01$0.01$3001M
GPT-5.6 SolOpenAI≈0$0$0.01$0.01$300—
Claude Opus 5Anthropic≈0$0$0.0125$0.0125$3751M
GPT-5.5OpenAI≈0$0$0.015$0.015$450—
Claude Fable 5.1Anthropic≈0$0$0.025$0.025$7501M
GPT-6 AstraOpenAI≈0$0$0.025$0.025$750—

≈ marks a count estimated with OpenAI’s o200k_base encoding, because that model’s own tokenizer cannot run in a browser — the real figure can differ, and Anthropic notes its newer models produce roughly 30% more tokens than earlier ones. Exact counts are shown without ≈. Prices verified 2026-09-24.

Your text is tokenized in this browser and never uploaded.

Claude Haiku 4.5 pricing at a glance

  • Provider: Anthropic
  • Input: $1.00 per million tokens
  • Output: $5.00 per million tokens
  • Cached input: $0.10 per million tokens
  • Context window: 200,000 tokens
  • Prices verified: September 24, 2026

About Claude Haiku 4.5

Haiku 4.5 is the low-cost Claude option at $1 per million input tokens and $5 per million output, with cache reads at $0.10 per million. It has a 200,000-token context window, smaller than the 1M windows of the newer Claude models, and it still uses the earlier Claude tokenizer, so token counts for the same text are lower than on Claude 4.7 and later.

What Claude Haiku 4.5 costs in practice

A call with 1,000 input tokens and 500 output tokens costs $0.0035 on Claude Haiku 4.5, so a thousand such calls cost $3.50. With prompt caching, a cache read is priced at $0.10 per million tokens, 90% below the standard input rate. Its 200k-token context window holds roughly 150,000 English words, or about 300 pages at 500 words a page; filling all of it once costs $0.20 in input tokens.

How this Claude Haiku 4.5 token counter counts tokens

Claude Haiku 4.5’s own tokenizer cannot be run in a browser, so this page counts with OpenAI’s o200k_base encoding and marks the result ≈. Use it to size a prompt and compare costs, then confirm against the usage figures the API returns.

How to use it

  1. Paste your prompt into the box, or drop a .txt or .md file on it.
  2. Set the expected output length and how many calls you make per day.
  3. Compare the cost per call and per month across models; the cheapest is highlighted.
  4. Tick the cache option if your prompt prefix is reused, then copy the table if you need it elsewhere.

Frequently asked questions

How many tokens is my text in Claude Haiku 4.5?

Paste it into the box above. The number shown for this model is an estimate made with o200k_base and marked ≈, so the API’s own count can differ.

How much does Claude Haiku 4.5 cost per million tokens?

$1.00 per million input tokens and $5.00 per million output tokens, with cached input at $0.10, as of September 24, 2026.

Is my prompt sent to a server?

No. Counting runs in a background thread in your browser. The site records only an anonymous counter such as how many models were compared, never the text you paste.

Prices verified September 24, 2026 from Anthropic’s pricing page. Confirm on the provider’s site before budgeting.