OpenAI introduces new version of ChatGPT
OpenAI presented the multimodal generative model GPT-4o at the Spring Update presentation. This AI is trained to work with text, photos, and videos. Communication with it will be available absolutely free and without registration, according to the internet edition about startups, innovations, and websites TechCrunch.
What's new in GPT-4o
OpenAI's Technical Director, Mihaela Murot, announced at the presentation at OpenAI's office in San Francisco that GPT-4o has the intelligence level of GPT-4, as well as improved abilities in working with text, images, and audio.
GPT-4o will significantly enhance the performance of the ChatGPT AI chatbot, which already supports a voice mode using text-to-speech conversion. The new model will bolster this feature, allowing users to interact with the chatbot as an assistant, including the ability to ask questions and interrupt responses.
GPT-4o provides real-time responsiveness and can recognize emotions in the user's voice, generating responses in different emotional styles.
Additionally, GPT-4o will enhance ChatGPT's visual capabilities, enabling quick responses to questions about presented photos or desktop screens.
From today, GPT-4o is available to both paid and free ChatGPT users. Subscribers to the ChatGPT Plus and Team plans will receive five times more messages.
Free users will also gain access to new features such as internet information search, data analysis, chart creation, image and file handling, as well as improved memory of previous interactions. Free users will have the opportunity to use the GPT Store.
GPT-4o supports more languages with improved performance in 50 languages. In the GPT-4o API, it works twice as fast as GPT-4 (including GPT-4 Turbo) and costs half as much for higher speed limits.
Currently, voice communication support in the GPT-4o API is not included for all clients but will be launched for a small group of trusted partners in the coming weeks due to the risk of misuse.