OpenAI held its long-awaited event today. The company introduced its new flagship model, GPT-4o, at the event. The model can talk, see and hear like a real person.
Pioneering the artificial intelligence revolution and leaving everyone speechless with its models OpenAIheld its event, which has been awaited for days, today. The company during the event ChatGPT chat bot and GPT-4 language model made important announcements, including upcoming innovations.
The most striking of today’s announcements is The company’s new flagship language model was GPT-4o. This model not only outperforms the existing GPT but is also much faster.
GPT-4o; Can reason through voice, text and image
The new GPT-4o model that the company will offer to its users will power the ChatGPT chat bot. The model, which is described as much more efficient and ahead of previous versions of GPT, will be able to reason through voice, text and image. According to the statements, GPT-4o, a built-in multi-model artificial intelligence model. This means that it can understand voice, text and image and produce content.
We can say that there has been a serious improvement, especially on the voice response side. Users can now model real-time with less lag, feeling much more realistic speeches can realize it. According to OpenAI, it can respond to sound in as little as 232 milliseconds. It’s almost as fast as talking to a human. Previously, delays in voice mode averaged 2.8 seconds.
In addition, you can even interrupt ChatGPT while he is answering and ask him to change his answer. For example, in the live demo at the event, OpenAI executives ask the model to tell a story about a robot. While the model is talking, they interrupt her and ask her to express different emotions. ChatGPT; By making this change instantly, he can fulfill the other person’s wishes. You can take a look at those moments in the video above.
The model’s built-in advanced visual capabilities were also demoed. The model can “see” and comment on things shown to her through the device’s camera. For example, in one demo, an equation written on paper was shown to the model and the model was asked for help in solving it. ChatGPT helped them find the solution. When “I Love You ChatGPT” was written on paper, it responded with an emotional voice, just like a human.
Can do real-time translation surprisingly well
*If you can’t view the video link here Click.
Another demo at the event was the translation capabilities of the new model. OpenAI demonstrated how GPT-4o can perform real-time translation. Mira Murati spoke in Italian while other OpenAI employees spoke in English. The model was able to translate the sentences and transfer them to the other side very quickly, with almost no delay.
It can read your screen and help with coding via the desktop application:
*If you can’t view the video link here Click.
Another live demo of GPT-4o showed off some of its coding capabilities. In this demo shown through ChatGPT’s new desktop application, the model was able to understand and analyze the written code. He was also able to explain what he saw.
GPT-4o can look at you through the camera and make comments:
*If you can’t view the video link here Click.
OpenAI shared videos of GPT-4o’s capabilities after the event ended. For example, in one of them, we can see the model’s conversation with a person via the front camera. ChatGPT can understand what that person looks like through the camera and make comments about their appearance. He even comments on the environment that person is in. It would not be wrong to say that it has become difficult to distinguish from humans.
In another demo, we can see that the user can understand their facial expressions and emotions by looking at their face. ” to himWith a smile on his face and a little excitement“You look so happy and cheerful.” says.
He even makes sarcastic jokes:
*If you can’t view the video link here Click.
In another example, the model is asked to make sarcastic jokes. GPT-4o can actually talk sarcastically and make jokes.
Here’s an example of how you can interrupt him and make him do what you want:
In this video, the model is asked to count to 10. The OpenAI employee across from him interrupts him after he starts counting and He asks her to count faster. He said he was able to fulfill all requests successfully, and even felt fed up at times. “Ok” We can even hear him say.
Two GPT-4os chatting, singing together
In a video, we can see GPT-4o chatting with another GPT-4o. They even sing in a duet towards the end of the video.
This is how the model reacts when she sees a dog:
It can be the “eye” of the visually impaired
In another example, we take a look at a very useful use of the model’s visual abilities. Here, the model explains the surroundings to a visually impaired individual and literally shows him/her. eye is happening.
GPT-4o will also be available to free users!
OpenAI also gave very good news about its language model. According to the company, GPT-4o Without paying any money via the free version of ChatGPT can be used. However, one message will be the limit. When this limit is exceeded, it will automatically revert to GPT-3.5. The company stated that GPT-4o’s text and image capabilities are available as of today. It was reported that the new sound mode will come to Plus users in the future.