OpenAI unveils GPT-4o, bringing man-machine interaction even closer together 🤖

Cliquez ici pour lire en français

OpenAI has just unveiled GPT-4o, a new and improved version of its generative artificial intelligence model. Nicknamed « o » for « omni », this system marks a further step towards more natural, fluid man-machine interaction.

Multimodal capabilities for natural interaction ✨

The GPT-4o’s main innovation lies in its multimodal capabilities. It accepts and generates combinations of text, audio, images and video, where previous models were generally limited to one modality at a time. This enables richer, more contextualized conversations. For example, a user could show a photo to GPT-4o and have a dialogue around the objects represented.

GPT-4o, integrated into the heart of ChatGPT, enables conversations where the user can talk to the assistant as if it were a human. With an average response time of 320 milliseconds, comparable to that of a human being, the model offers near-instant interaction. This responsiveness, coupled with voice enhancements with realistic modulations, delivers a much more fluid and natural conversational experience.

GPT-4o also brings significant advances in terms of multilingual comprehension. Exceeding the limited capabilities of GPT-4 for English, it shows significantly improved performance for the processing of languages other than English, both written and spoken. In particular, OpenAI has demonstrated real-time translation capabilities with ChatGPT.

GPT-4o capabilities are already deployed in ChatGPT, OpenAI’s chat application. Paying users benefit from early access to multimodal functionalities. Free users can take advantage of GPT-4-level intelligence, data analysis and graphing, photo chats and much more.

New interface and applications for desktops💻

OpenAI has also announced a redesign of ChatGPT’s interface, making it more user-friendly and conversational.

Source : ChatGPT

A macOS application is launched, integrating GPT-4o’s new capabilities. It enables text conversations, and in the future, audio and video interactions will be possible thanks to GPT-4o’s advanced capabilities. Voice mode, already integrated, will soon be enhanced for even faster, more natural interactions. Users can already initiate voice conversations by clicking on the headset icon in the desktop application.

Users can interact with ChatGPT via a simple keyboard shortcut (Option + Space) to ask questions or add attachments without opening the dedicated application. This feature will be available to all users, paid and free, with a Windows version in development.

Do you use ChatGPT on a daily basis? What do you think of these new features? Tell us in the comments section.