TokenSave logo TokenSave
100% in-browser · Nothing is uploaded

OpenAI Token Counter & GPT API Cost Calculator

Count tokens and see what GPT-6.2 Astra, GPT-6.2 Sol, GPT-6 Sol and Luna charge for your prompt, in any language.

The same text can use from 3% (Chinese) to 144% (Punjabi) more tokens than English — see how we measured 41 languages.

How we measured →

🔍 See how it splits into tokens

Each colour is one token (GPT-4o and later). Boxes with a dashed outline are a single character that takes several tokens.

Colours just split tokens. Same price each: more boxes = higher cost.

Official pricing & tokenizer references:

How do I count tokens for GPT-6?

Paste your text above with the GPT tab selected. You get the token count, characters and the input and output cost on each GPT model. OpenAI has not published GPT-6's tokenizer yet, so GPT-6 counts are an estimate based on o200k; GPT-4o, GPT-4.1 and GPT-5 mini counts are exact.

Is this the same tokenizer as tiktoken?

Yes for the exact models: it uses the same o200k encoding that OpenAI's tiktoken library uses for GPT-4o and GPT-5 mini, running in your browser.

How much does GPT-6 cost per token?

Per million input / output tokens: GPT-6.2 Astra $8 / $40; GPT-6.2 Sol $2 / $10; GPT-6 Astra (legacy) $10 / $50; GPT-6.1 Sol $2 / $10; GPT-6.1 Terra $1 / $5; GPT-6 Sol $2 / $10; GPT-6 Luna $0.1 / $0.5; GPT-5 mini $0.25 / $2; GPT-4o $2.5 / $10.

Should I use ChatGPT Plus or the API?

It depends on how much you use. The ChatGPT Plus or API calculator on this site compares a monthly subscription with pay-per-token API cost for your own usage.

Is my text sent anywhere?

No. The count runs in your browser and your text is never uploaded.

How GPT counts tokens

OpenAI publishes its tokenizer, so GPT counts can be exact. GPT-4o, GPT-4.1 and GPT-5 mini and nano use the o200k tokenizer, and this counter uses the same one, so their numbers match what OpenAI bills. OpenAI has not published the tokenizer for GPT-5.4 and later, including GPT-6, so for those models the count is an o200k-based estimate.

The biggest price decision on OpenAI is the model, not a discount. GPT-6.2 Astra costs 300% more than GPT-6.2 Sol per token, and Sol costs 1,900% more than GPT-6 Luna. Most apps run well on Sol or Luna and send only the hard requests to Astra.

Two ways to pay less on the same model: prompt caching bills repeated input (fixed instructions, documents) at a tenth of the normal input price, and the Batch API charges about half for work that can wait up to 24 hours. Requests with more than 272,000 input tokens are billed at a higher rate than the table below.

OpenAI API prices

ModelInput / output per 1M1,000 chat requests
GPT-6.2 Astra$8 / $40$28.00
GPT-6.2 Sol$2 / $10$7.00
GPT-6 Astra (legacy)$10 / $50$35.00
GPT-6.1 Sol$2 / $10$7.00
GPT-6.1 Terra$1 / $5$3.50
GPT-6 Sol$2 / $10$7.00
GPT-6 Luna$0.1 / $0.5$0.35
GPT-5 mini$0.25 / $2$1.18
GPT-4o$2.5 / $10$7.75

Chat request = 1,500 input + 400 output tokens of English text. Standard list prices, no caching or batch discounts. (2026-10-14)

More on GPT costs