New ChatGPT-4.5 Released CEO Altman Says: It’s Also Emotionally Intelligent On Thursday, OpenAI unveiled its largest and most powerful AI model to date, GPT-4.5. The company’s CEO Sam Altman described it in a post on X as “the first model that feels like you’re talking to a thinking person.”
Altman also described the model as “huge” and “expensive.” OpenAI had planned to release it on Plus and Pro at the same time, but that was reportedly not possible due to a lack of GPUs.
Altman added, “We’ll be adding tens of thousands of GPUs next week and then rolling it out to the Plus tier.”
On Thursday, the company released a research preview of GPT-4.5 to users and developers on the API who pay $200 for ChatGPT Pro. Next week, OpenAI said it will bring it to ChatGPT Plus, Teams, and EDU users.
Its Oven (o1) series of reasoning models (released last year) are designed to think before they respond, which is more useful for quantitative tasks.
GPT-4.5 is ahead of all other OpenAI models in terms of speaking naturally and acting emotionally intelligently. Its broad knowledge base and strong contextual understanding allow it to respond more efficiently to social cues than OpenAI’s STEAM-focused reasoning model Oven.
GPT-4.5 is more efficient at rendering content written in angry language than O One. GPT-4.5 is designed to be a “good assistant.” Conversations with it are said to feel “warmer, more comfortable, and emotionally nuanced.”
According to Open AI, the model’s “knowledge base, strong attunement to user intent, and improved emotional intelligence make it well-suited for tasks like writing, programming, and practical problem-solving.”
