AI Integration

Models

Currently App Suite supports 3 different Open AI models: GPT-3.5 Turbo, GPT-4 and GPT-4o

Models differ in, response time: newer ones are generally faster, context length: newer models can work with longer texts and have longer 'short term memory' training data: newer models are trained with more recent data feature set: accepted input types, language support, etc pricing: models are paid per 1 million tokens used

The following table shortly explains the differences of each model:

You can get more details about the models here: https://platform.openai.com/docs/models

Pricing Example

Pricing is done per million tokens. This includes used input and output tokens. Used output tokens are more expensive.

Here is an example for a translation task (prices checked 8th July 2024):

Important meeting with customer next thursday. Don't be late.

Greetings
Sam

translated to German: ``` Wichtiges Treffen mit Kunde nächsten Donnerstag. Bitte sei nicht zu spät.

Grüße Sam ``` The request counts the following tokens:

prompt tokens: 26 completion token: 20 total tokens: 46