OpenAI debuts GPT-4o that can laugh, sing, teach math and more

OpenAI has unleashed a wave of innovation with the launch of GPT-4o, a seemingly natural and real-sounding voice assistant with vision capabilities. The unveiling occurred during the OpenAI Spring Update event, presided over by OpenAI’s CTO, Mira Murati, who heralded GPT-4o as a groundbreaking advancement that democratises the cutting-edge capabilities of GPT-4.

With the introduction of GPT-4o, free users of ChatGPT are poised to get their hands on the new tech, that gives them access to GPT-4's capabilities, which have been behind a paywall so far.

Central to this update is the debut of ChatGPT’s desktop application, alongside a revitalised web UI. These enhancements are designed to streamline access to ChatGPT and foster more intuitive user experiences.

OpenAI revealed that GPT-4o’s capabilities will be progressively rolled out, starting with text and image functionalities integrated into ChatGPT. Notably, the new model boasts the ability to emulate human cadences in verbal responses and even discern users' moods, evoking parallels with the emotive AI depicted in Spike Jonze’s acclaimed film "Her."

According to OpenAI, GPT-4o operates with unprecedented speed and agility, facilitating real-time reasoning across text, audio, and video formats. This advanced iteration will power OpenAI's flagship ChatGPT chatbot, promising enhanced user experiences across the board.

Excitement abounds as OpenAI prepares to make GPT-4o available to all users in the coming weeks, extending its reach to both paid subscribers and those utilising the free version of ChatGPT. Additionally, OpenAI unveiled a series of videos showcasing GPT-4o in action.

These include a video of a man making the AI guess what he was up to by just looking at his surrounding.

In another film, two humans ask the AI to commentate a game of Rock, Paper, Scissors, while also making it announce the winners.

In another film, a user asks ChatGPT to become his translator. The AI assistant instantly helps him engage in a real life conversation with an Italian-speaking woman, with real-time translation.

There are several other films that the company has unveiled that depict other features and use cases of the new GPT-4o, on the company official YouTube channel.