OpenAI token counter

Paste a prompt, document or code to see its token count and what it costs as input and as output on each OpenAI model. Counting runs in your browser.

Accuracy: Exact for the text itself on GPT-5.x models: tiktoken maps every gpt-5 model name to the o200k_base encoding, and this page runs that encoding in your browser. Two limits: the API adds a few formatting tokens for message roles and boundaries, and OpenAI has not published the encoding for GPT-6 models (they are not in tiktoken's model map), so for those the o200k_base count is the closest public reference, not a confirmed exact count.

The exact OpenAI tokenizer (about 1 MB) loads when you start typing.

Cost of this text on OpenAI models

ModelTokensAs inputAs output× 1,000

Standard prices, no caching or batch discount. "As output" is what the same number of tokens costs if the model generates it. Long-context rates apply automatically when the text crosses a model's threshold. Full requests with caching and agent loops: calculator.

OpenAI tokenizer facts

TokenizerPublic, in the open-source tiktoken library. o200k_base: gpt-5*, gpt-4.1, gpt-4.5, gpt-4o, o1, o3, o4-mini. cl100k_base: gpt-4, gpt-3.5-turbo and the text-embedding-3 models. o200k_harmony: gpt-oss. Source
GPT-6 modelsNot listed in tiktoken's model-to-encoding map. OpenAI has not published which encoding GPT-6 models use. Source
Exact count via APIPOST /v1/responses/input_tokens takes the same input as the Responses API (text, messages, images, files, tools) and returns input_tokens. The count includes formatting tokens for message roles and boundaries that a local tokenizer does not see. Source
Rule of thumbAbout 4 characters or 0.75 words per token for English text. Source

Checked against the linked official documentation on 2026-10-30.

Get the exact count from the official API

Python (tiktoken, local)

import tiktoken

enc = tiktoken.get_encoding("o200k_base")
print(len(enc.encode("Tell me a joke.")))

cURL (API, includes formatting tokens)

curl https://api.openai.com/v1/responses/input_tokens \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-6-astra",
    "input": "Tell me a joke."
  }'
# → {"object": "response.input_tokens", "input_tokens": ...}

Use tiktoken for fast local counts of plain text on GPT-5.x models, and the input_tokens endpoint when you need the number the API will bill for a full request, or for GPT-6 models. Official documentation

Which OpenAI models this page counts exactly

ModelCount on this page
GPT-6 Astrao200k_base reference
GPT-6.1 Solo200k_base reference
GPT-6.1 Terrao200k_base reference
GPT-6 Lunao200k_base reference
GPT-6 Solo200k_base reference
GPT-5.6 Solexact (o200k_base)
GPT-5.6 Terraexact (o200k_base)
GPT-5.6 Lunaexact (o200k_base)
GPT-5.5exact (o200k_base)
GPT-5.4exact (o200k_base)
GPT-5.4 miniexact (o200k_base)

6 of 11 models have a published encoding. For the rest, use the input_tokens endpoint when the number matters.

What OpenAI tokens cost

Cost of a given number of billed tokens on each model at standard rates (prices verified 2026-11-17):

Model1K in1K out10K in10K out100K in100K outPer 1M in / out
GPT-6 Astra$0.01$0.05$0.10$0.50$1.00$5.00$10.00 / $50.00
GPT-6.1 Sol$0.0020$0.01$0.02$0.10$0.20$1.00$2.00 / $10.00
GPT-6.1 Terra$0.0010$0.0050$0.01$0.05$0.10$0.50$1.00 / $5.00
GPT-6 Luna$0.00010$0.00050$0.0010$0.0050$0.01$0.05$0.10 / $0.50
GPT-6 Sol$0.0020$0.01$0.02$0.10$0.20$1.00$2.00 / $10.00
GPT-5.6 Sol$0.0040$0.02$0.04$0.20$0.40$2.00$4.00 / $20.00
GPT-5.6 Terra$0.0020$0.01$0.02$0.12$0.20$1.20$2.00 / $12.00
GPT-5.6 Luna$0.00020$0.0012$0.0020$0.01$0.02$0.12$0.20 / $1.20
GPT-5.5$0.0050$0.03$0.05$0.30$0.50$3.00$5.00 / $30.00
GPT-5.4$0.0025$0.01$0.03$0.15$0.25$1.50$2.50 / $15.00
GPT-5.4 mini$0.00075$0.0045$0.0075$0.04$0.07$0.45$0.75 / $4.50

Long-context pricing: GPT-6 Astra, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Luna, GPT-6 Sol, GPT-5.6 Sol, GPT-5.6 Terra, GPT-5.6 Luna, GPT-5.5, GPT-5.4 bill the whole request at 2× input and 1.5× output once the prompt exceeds 272,000 tokens. The output columns above assume a short prompt.

Per-model details: GPT-6 Astra, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Luna, GPT-6 Sol, GPT-5.6 Sol, GPT-5.6 Terra, GPT-5.6 Luna, GPT-5.5, GPT-5.4, GPT-5.4 mini. Compare with other providers in the pricing table.

FAQ

Is this OpenAI token count exact?

For the text itself on GPT-5.x models, yes: tiktoken maps those models to o200k_base and this page runs that encoding (gpt-tokenizer, an open-source port). The API adds a few tokens for message roles and boundaries. GPT-6 models are not in tiktoken's model map, so for GPT-6 Astra, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Luna, GPT-6 Sol the count is the closest public reference, not a confirmed exact count.

Which tokenizer does each OpenAI model use?

Per tiktoken's model map: o200k_base for gpt-5 models, gpt-4.1, gpt-4.5, gpt-4o, o1, o3 and o4-mini; cl100k_base for gpt-4, gpt-3.5-turbo and the text-embedding-3 models; o200k_harmony for gpt-oss. GPT-6 models are not listed.

How do I get the exact token count from the OpenAI API?

POST your request to /v1/responses/input_tokens. It accepts the same input as the Responses API (text, messages, images, files, tools) and returns input_tokens, including the formatting tokens a local tokenizer cannot see.

How many tokens is a word?

OpenAI's rule of thumb for English is about 4 characters or 0.75 words per token, so 1,000 words is about 1,333 tokens. Code and non-English text use more tokens per word.

How much do 100,000 tokens cost on GPT-6.1 Sol?

$0.20 as input and $1.00 as output at standard rates ($2.00 / $10.00 per 1M tokens, verified 2026-11-17). Cached input and the batch API are cheaper; see the calculator.

Is my text sent to OpenAI?

No. The tokenizer runs in your browser and nothing you paste leaves the page.

Other token counters

Tokenizer: gpt-tokenizer 3.4.0 (MIT, license), served from this site.