ChatGPT and GPT-4 are quite cheap for content creation (yes, even though GPT-4 has become more expensive again). With my interactive calculator you can find out what costs you will incur when using the API for content creation with ChatGPT, GPT-3.5 or GPT-4. This way you always keep track of the costs of your artificial intelligence and can plan your budget more effectively.
In addition, the calculator was created 100% with ChatGPT. So not being able to program (like me) is no longer an excuse.
How are OpenAI costs calculated?
The calculator uses the information provided on the official OpenAI website to calculate the estimated price based on the number of words. The calculator assumes that 1000 tokens are equal to about 750 words, as recommended by OpenAI.
For all models, the costs are for the prompt (i.e. your input) and hungary number dataset the output created. However, this is different with GPT-4 . Here, for the first time, there are different prices for the prompt and the result. However, my calculator takes this into account.
Important: The GPT-4 API is currently only available via a waiting list, but I have already included the prices.
Interactive Calculator for OpenAI API
Simply enter the amount of words you want to generate or select one of the various examples to get a feel for how cheap GPT-4, GPT-3.5 and ChatGPT are. The calculator will then show you the total price you need for the API to generate that amount of text below.
How many words do you want to create?
number of words
Estimated price:
NaN words cost about $NaN (including 100 words for the prompt)
This calculator is made entirely with ChatGPT and inspired by this calculator .
What is a token?
You can think of tokens as parts of words used for natural language processing. For English texts, 1 token is about 4 characters or 0.75 words. For comparison, the collected works of Shakespeare contain about 900,000 words or 1.2 million tokens.
OpenAI has also released a tokenizer tool where you can test this ratio. The amount of tokens is also displayed in the playground.
Which model should you choose?
While Davinci or ChatGPT is generally the most powerful model, the other models can perform certain tasks extremely well and in some cases significantly faster. They also have cost advantages. For example, Curie can perform many of the same tasks as Davinci, but faster and at a tenth of the cost. The GPT-4 API is currently only available via a waitlist.
Also note that with appropriate fine-tuning of the models, you can also use cheaper models from OpenAI really effectively. Depending on the project and volume, you can save a lot of money here, but get the same results.