A new neuronics model GPT-4o has been unveiled

On the evening of May 13, OpenAI held its Spring Update presentation, where it unveiled new products. The main focus was on the new flagship language model, as well as a number of smaller announcements.

ChatGPT will get a desktop application, so far only a Mac version has been confirmed. The app aims to make it easier to integrate the chatbot into users' workflows. The ChatGPT interface has been updated, making the web version more user-friendly and easier to use. However, there were no significant changes.

A new model, the GPT-4o, was announced. This is OpenAI's multimodal flagship model, capable of handling text, audio and video. For example, you can have a voice dialog with it, controlling the model's speech style in real time. Unlike previous OpenAI developments, GPT-4o is a single model rather than a collection of several. It supports 50 languages and is able to maintain a live conversation without delayed responses.

GPT-4o can analyze in real time not only static images, but also video, for example, from a smartphone camera. The model can also be used on the desktop to analyze code in an adjacent editor (code must be manually highlighted).

GPT-4o will be free for all OpenAI users. Access to the new model will be made available gradually over several weeks.