OpenAI presents its new ChatGPT-4o
OpenAI has announced a new artificial intelligence model, called GPT-4o, and it promises to be a new leap in the evolution of ChatGPT and its capabilities. Specifically, OpenAI promises a much faster and smarter ChatGPT, improving its capabilities in text, audio and video, including almost immediate and much more natural responses.
Naturalness, Vision and Impact
The presentation was led by Mira Murati, CTO of OpenAI. Murati started the event by announcing the new GPT-4o multimodal model. In the case of the new model, it has intelligence at the level of GPT-4 but is much faster, the latency of responses is now minimal.
Naturalness and speed
The chatbot can converse with the user in a more natural and faster way, without us having to wait several seconds before it responds to us, as was necessary until now. And if that wasn't enough, OpenAI has also surprised us by revealing that ChatGPT can now also be used as a real-time translator. According to OpenAI, GPT-4o has a voice response time of about 320 milliseconds, similar to that of human beings.
Real-time translation
One of the most impressive new features they demonstrated on stage was the ability to translate a conversation in real time. Murati had a conversation with an OpenAI engineer, she speaking in Italian and he in English, and immediately ChatGPT could translate the conversation in a natural way, as if a third person was helping these two to communicate. He responded naturally and immediately to queries made during the event.
New vision capabilities
The new model also improves its vision capabilities, which means that we can show it an image or a photo for it to analyze and provide us with information about it. In addition, in another demonstration in which OpenAI had a chatbot talk to another chatbot, the AI used the mobile camera to analyze the elements present in a room and the user's clothing.
When will it be available?
The new GPT-4o capabilities in text and images start to be available in ChatGPT from today, and will progressively reach all users, both for people subscribed to ChatGPT Plus and for those using the free version of ChatGPT.
The enhanced GPT-40 conversation will be coming within the next few weeks to ChatGPT Plus users. OpenAI will announce more details of its new multi-modal model soon.