- Openai launched GPT-4.5, but only for Chatgpt Pro users at the moment. In addition, users get it next week
- GPT-4.5 is supposed to be more intelligent and more intuitive emotionally than the models of anterior AI but lack deep reasoning power
- OPENAI claims that GPT-4.5 will provide better conversations and fewer hallucinations than ever
Openai officially launched GPT-4.5, the most advanced version of its AI models, but it has no deep reasoning capacity, because it is designed to be a versatile model. The latest model is available on Chatgpt, but for the moment, you will have to be a professional subscriber to access it.
The model is deployed as a “research overview”, which means that it is imperfect, and Openai hopes that users will help sort the bugs or clumsiness. Despite this, the company affirms that the first testers report that the GPT-4.5 feels more natural, intuitive and emotionally intelligent. This is a welcome upgrade for those of us who have already asked Chatgpt for information and who had the impression that the answer was only an obsolete Wikipedia page.
GPT-4.5 is supposed to be even better to understand the context and recognize the social clues than its predecessors. The overall feeling is supposed to be quite indistinguishable to discuss with a real person. He certainly seems to have impressed Openai CEO, Sam Altman, who called him “the first model that feels like talking to a thoughtful person” in an article on X. Altman added that GPT-4.5 will come to Chatgpt Plus and the team subscribers next week, but you will have to pay $ 200 per month to play with for the moment.
GPT-4.5 is ready! Good news: this is the first model that wants to tell me about a thoughtful person. I had several moments where I sat on my chair and I was amazed to get good advice from a new AI.BAD: it is a giant and expensive model. We…February 27, 2025
Fewer hallucinations, more empathy
In particular, GPT-4.5 is not only supposedly more intelligent and more creative, but it should also produce less hallucinations. This last part is particularly important because hallucinations were often a point of collision for provisional users of Chatgpt who did not want to count on a chatbot which could confidently invent things. In terms of performance, GPT-4.5 is supposed to be a big step in place and really good with general knowledge. However, this does not mean that it can manage the coding and problem solving as well as the O1 and O3 models of Openai or rival options like R1 of Deepseek.
Instead, GPT-4.5 specializes in empathy, with a high emotional quotient (EQ). The AI still cannot feel things, but the GPT-4.5 is supposed to be better to resume the nuance. If you talk about a bad day, it won’t just answer: “I’m sorry you feel like that.” Instead, it is designed to offer more thoughtful answers.
“By vibes, really means the equalizer of the model, how much he feels he feels and how warm his tone is,” said Openai Rapha Gontijo engineer in a live demonstration by GPT-4.5. “We have measured this by selecting a set of opinion prompts and selecting our trainers for those who line up with our vibrations.”
If you are a chatgpt pro user, you will be among the first to see if GPT-4.5 really has the vibrations of a human and up to the media threw. If you don’t spend as much, you will have a chance early enough. Anyway, with Amazon announcing an Alexa + Plus talkative, yesterday, fueled by AI, the AI arms race which has just warmed up, and this time, it seems that chatbots will be able to imitate the way you feel more realistic than ever.
