20 minutes chat

The price for the one-time chat access is €5.49 Please click on the button next to the desired organisation to pay. You will then be redirected to PayPal. You do not need a PayPal account as it is also possible to pay as a guest. After successful payment you will be redirected to the chat. The chat is set up for a one-time use of about 20000 tokens (see description of tokens) for 20 minutes. A counter shows the usage.

By using the chat, you agree to our privacy policy and terms and conditions. No personal data is collected except during the payment process via PayPal.

Tokens are the building blocks that make up texts in AI language processing. A Token can be a word, part of a word or a character (e.g. a full stop or comma). The exact number of tokens depends on the text.

Example:

The sentence "Hello, how are you?" consists of 6 tokens:
- "Hello, „,“, "like", "goes", "it", "You?".

AI models such as GPT process texts as sequences of tokens. There are limits to the number of tokens that a model can process in a request (input + response). With GPT-4 for example, the maximum number of tokens is 4096 Tokens.

Why are tokens important?

ChargesThe utilisation of a model is calculated based on the number of tokens.
PerformanceToo many tokens can lead to API errors if the limit is exceeded.
Shortening messagesWhen the limit is reached, older tokens must be removed to make room for new ones.

In short, tokens are a technical way of breaking down texts into smaller units that can be processed by AI models.

The cost of using a model such as GPT-4 (e.g. via the OpenAI API) is made up of several factors. Here is an overview of how the costs are calculated:

1. number of tokens

Tokens are calculated for entries and responses:
- Input tokensAll tokens that are sent to the model (e.g. system message, message history, user input).
- Response tokens: All tokens that the model generates in response.

Example:

You send a request with 100 tokens, and the model responds with 50 tokens.
Total cost = 150 tokens.

2. price per 1,000 tokens

OpenAI calculates the costs based on the number of tokens used per model

3rd model version

Costs vary depending on the model:
- GPT-4 is more expensive, but delivers better results.
- GPT-3.5 is cheaper and faster, but less powerful.

4. optimisation of the request

The length of the system message and the message history influence the total cost:
- Shorter messages = fewer tokens = lower costs.
- Longer conversations can be optimised by shortening old messages.

Formula for cost calculation

`Total costs = (input tokens × input price/1,000) + (response tokens × response price/1,000)`

Additional considerations

API limits:
- OpenAI sets monthly limits for tokens based on your subscription.
Network costs:
- Although low, API calls themselves can incur additional costs (e.g. data transfer).
Administrative costs:
- If you host an application, you will incur server costs or infrastructure fees (e.g. for the web server on which your tool runs).

To summarise: The costs depend on the number of tokens processed (input + response), the model used and the length of the requests and responses. Optimisations such as shorter messages and the right model can help to reduce costs.

20 minutes chat

By using the chat, you agree to our privacy policy and terms and conditions. No personal data is collected except during the payment process via PayPal.

Tokens are the building blocks that make up texts in AI language processing. A Token can be a word, part of a word or a character (e.g. a full stop or comma). The exact number of tokens depends on the text.

Example:

The sentence "Hello, how are you?" consists of 6 tokens:

"Hello, „,“, "like", "goes", "it", "You?".

AI models such as GPT process texts as sequences of tokens. There are limits to the number of tokens that a model can process in a request (input + response). With GPT-4 for example, the maximum number of tokens is 4096 Tokens.

Why are tokens important?

ChargesThe utilisation of a model is calculated based on the number of tokens.

PerformanceToo many tokens can lead to API errors if the limit is exceeded.

Shortening messagesWhen the limit is reached, older tokens must be removed to make room for new ones.

In short, tokens are a technical way of breaking down texts into smaller units that can be processed by AI models.

The cost of using a model such as GPT-4 (e.g. via the OpenAI API) is made up of several factors. Here is an overview of how the costs are calculated:

1. number of tokens

Tokens are calculated for entries and responses:

Input tokensAll tokens that are sent to the model (e.g. system message, message history, user input).

Response tokens: All tokens that the model generates in response.

Example:

You send a request with 100 tokens, and the model responds with 50 tokens.

Total cost = 150 tokens.

2. price per 1,000 tokens

OpenAI calculates the costs based on the number of tokens used per model

3rd model version

Costs vary depending on the model:

GPT-4 is more expensive, but delivers better results.

GPT-3.5 is cheaper and faster, but less powerful.

4. optimisation of the request

The length of the system message and the message history influence the total cost:

Shorter messages = fewer tokens = lower costs.

Longer conversations can be optimised by shortening old messages.

Formula for cost calculation

`Total costs = (input tokens × input price/1,000) + (response tokens × response price/1,000)`

Additional considerations

API limits:

OpenAI sets monthly limits for tokens based on your subscription.

Network costs:

Although low, API calls themselves can incur additional costs (e.g. data transfer).

Administrative costs:

If you host an application, you will incur server costs or infrastructure fees (e.g. for the web server on which your tool runs).

To summarise: The costs depend on the number of tokens processed (input + response), the model used and the length of the requests and responses. Optimisations such as shorter messages and the right model can help to reduce costs.