At a recent conference for IT professionals, research company OpenAI demonstrated its latest achievements in the field of artificial intelligence, introducing an updated GPT-4 Turbo model. This version is claimed to be more advanced and cost-effective than its predecessor, GPT-4.

Main characteristics and advantages of the GPT-4 Turbo:

  • Possibility of using an extended prompt. Users can formulate queries up to 128,000 tokens, which is equivalent to 365 pages of text.
  • The ChatGPT chatbot issues responses in a structured JSON format. This expands the possibilities of using the model in various applications and services.
  • A function for uploading user documents (in particular in PDF format), making it easier to work with your own content.
  • The updated model knowledge base, containing information as of April 2023, ensures greater relevance and reliability of the generated answers.
  • Cost-effectiveness of using APIs. The cost of requests in the new model is reduced by three times, and the generation of responses is reduced by half compared to the previous version.
  • Development of personalized chatbots (GPTs). Thanks to a simple constructor in the ChatGPT interface, which does not require special programming knowledge, a wide range of users will have access to it.
  • Own application store. It gives users the opportunity to share their GPTs and earn income from their use by others.
  • Text-to-speech functionality via the speech synthesis API. There are six different naturalistic human voices to choose from.

Currently, GPT-4 Turbo has two versions. One specializes in text analysis, while the other can recognize and understand the context of both text and images. OpenAI plans to make both versions publicly available in the coming weeks.