OpenAI has introduced a new AI model called GPTo (General Pre-trained Transformer-Omni) which improves on their previous ChatGPT model.
GPTo can reason across text, audio and video in real time, allowing it to understand and respond to different media formats simultaneously.
In a demonstration, GPTo had a voice conversation, helped with a math problem without immediately providing the answer, and assisted with a coding problem by looking at code on a screen.
It also attempted to detect mood from a selfie video and translated between English and Italian.
OpenAI says GPTo works faster than previous versions and will power the ChatGPT chatbot interface. It will be available to both free and paid users in the coming weeks.
Some analysts say OpenAI appears to be playing catch-up to competitors like Google in capabilities for conversational AI that can understand different media formats. They will reveal more at their upcoming I/O conference.
Source: AP
