\n<\/p>\n
OpenAI is releasing a new flagship generative AI model called GPT-4o, set to roll out \u201citeratively\u201d across the company\u2019s developer and consumer-facing products over the next few weeks.<\/p>\n
OpenAI CTO Muri Murati said that GPT-4o provides \u201cGPT-4-level\u201d intelligence but improves on GPT-4\u2019s capabilities across text and vision as well as audio.<\/p>\n
\u201cGPT-4o reasons across voice, text and vision,\u201d Murati said during a keynote presentation at OpenAI\u2019s offices. \u201cAnd this is incredibly important, because we\u2019re looking at the future of interaction between ourselves and machines.\u201d<\/p>\n
GPT-4, OpenAI\u2019s previous leading model, was trained a combination of both images and text, and could analyze images and text to accomplish tasks like extracting text from images or even describing the content of those images. But GPT-4o adds speech to the mix.<\/p>\n
What does this enable, concretely? A variety of things.\u00a0<\/p>\n
GPT-4o greatly improves the ChatGPT experience \u2014 ChatGPT being OpenAI\u2019s viral AI-powered chatbot. ChatGPT has long offered a voice mode that transcribes text from ChatGPT using a text-to-speech model. But GPT-4o supercharges this, allowing users to interact with ChatGPT more like an assistant.\u00a0<\/p>\n
For example, users can ask ChatGPT \u2014 powered by GPT-4o \u2014 a question and interrupt ChatGPT while it\u2019s answering. The model delivers \u201creal time\u201d responsiveness, OpenAI says, and can even pick up on the emotion in a user\u2019s voice, and generate voice in \u201ca range of different emotive styles.\u201d<\/p>\n
GPT-4o improves ChatGPT\u2019s vision capabilities as well. Given a photo \u2014 or a desktop screen \u2014 ChatGPT can now quickly answer related questions, from things like \u201cWhat\u2019s going on in this software code\u201d to \u201cWhat brand of shirt is this person wearing?\u201d<\/p>\n
In other news, OpenAI is releasing desktop version of ChatGPT and a refreshed UI.<\/p>\n
\u201cWe know that these models [are getting] more and more complex, but we want the experience of interaction to actually become more natural, easy, and for you not to focus on the UI at all, but just focus on the collaboration with [GPTs],\u201d Murati said.<\/p>\n