GPT-4o

GPT-4o ("o" for "omni") is a step towards a more natural human-computer interaction. It 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 as little as 232 milliseconds, and an average of 320 milliseconds, which is similar to the response time of a human in a conversation. It matches the GPT-4 Turbo's performance for English and code-based text, while significantly improving the performance for non-English text, as well as being much faster and 50 % cheaper in API. Compared to existing models, GPT-4o performs particularly well in vision and audio comprehension.