OpenAI has announced what it claims to be the next big leap in its AI tech endeavours. This comes in the form of the latest AI model, dubbed GPT-4o, with the letter “o” standing for omni, representing its ability to accept “any combination of text, audio, and image”. Similarly, it can also output in any combination of the aforementioned media, and “in as little as 232 milliseconds”.
The announcement also comes with a demo of a verbal conversation being had with the OpenAI GPT-4o. Similar to the way “conversations” work with chatbots, it looks like a person’s one-line question will still get a multi-sentence response. But probably the biggest difference with the new AI model is that you can interrupt it while it is rambling, and it will discard its remaining script and start taking in new info.
Say hello to GPT-4o, our new flagship model which can reason across audio, vision, and text in real time: https://t.co/MYHZB79UqN
Text and image input rolling out today in API and ChatGPT with voice and video in the coming weeks. pic.twitter.com/uuthKZyzYx
— OpenAI (@OpenAI) May 13, 2024
Of course, being a demo, things are going about as smoothly as things can get. The voices also sound natural, with believable change in intonation, and even pauses, almost like another person. It even gets sarcasm. Of course, the dead giveaway being its very rambly nature. The OpenAI demo also shows what appears to be two instances of GPT-4o speaking to each other and singing an impromptu song, with both singing alternate lines.
This variance in voice tone carries over to the other demos which sees the GPT-4o model comment on a person’s appearance based on what a phone’s front-facing camera is seeing. It can also be made to teach math and Spanish, as well as do live translation between English and Italian.
While it’s all very impressive sounding, it wouldn’t mean much if it was all out of reach. The good news is that it won’t be completely inaccessible, as OpenAI says that it will be rolling out GPT-4o to everyone, even free users of ChatGPT. Naturally, free users have usage limits. Beyond that, the company also announced that it is releasing a macOS version of the ChatGPT app for both free and premium users. It’s not exactly iOS 18-related, so it remains to be seen if talks for that is still ongoing. A Windows version is on the way, but with no specific date besides a “later this year” window.
(Source: OpenAI [1], [2], [3])
Follow us on Instagram, Facebook, Twitter or Telegram for more updates and breaking news.