O is for omni: Takes any combination of text, audio, and image as input and generates any combination of text, audio, and image as output. It can respond to audio inputs in just 232 milliseconds, averaging 320 milliseconds, which is similar to the human response time (opens in a new window) in a conversation.
Compared to the previous model, 4.0 Turbo, it is twice as fast, costs OpenAI half as much thanks to the efficiencies achieved (and this will allow it to reduce prices and extend some functions to free users) and has five times higher frequency limits (a parameter which indicates how often users can make requests to the model, for example, to generate text, analyze data, or interact in other ways).
Very useful parameter for developers and companies using the OpenAI API: they can now make more requests in less time, improving efficiency and allowing more intensive use of artificial intelligence in their applications without quickly reaching usage limits which may slow down or interrupt their services.
The new model also improved the quality and speed of ChatGpt for 50 different languages and will also be available through OpenAI’s API, so developers can start building applications using the new model today, Murati said.
Also shown was a real-time translation, from English to Italian, which Murati appears to be able to speak. The artificial voice in our language is much less natural than the one in English, but it worked.