OpenAI unveils GPT-4.5 AI model with better 'EQ'

With its advancements in Emotional Intelligence (EQ) and Natural Language Processing (NLP), GPT-4.5 is set to transform how AI models engage in more human-like, collaborative interactions, offering a deeper, more intuitive user experience in a variety of fields

OpenAI has introduced the GPT-4.5 AI model, the latest and most advanced version of its language model. Currently available as a research preview for ChatGPT Pro users and developers worldwide, GPT-4.5 offers significant advancements in Emotional Intelligence (EQ), Natural Language Processing (NLP), and overall user interaction, marking a leap forward in AI capabilities.

Breakthrough features in GPT-4.5

The GPT-4.5 AI model is designed to provide more accurate and intuitive responses, making it particularly adept at tasks like writing assistance, programming, and problem-solving. The model boasts an expanded knowledge base and improved alignment with user intent, allowing for more human-like interactions. OpenAI also highlighted that GPT-4.5 reduces the frequency of "hallucinations," where AI generates inaccurate or nonsensical information, further enhancing reliability.

OpenAI emphasized that while GPT-4.5 represents a notable enhancement in Emotional Intelligence (EQ), it does not introduce entirely new capabilities compared to earlier models like o1 or o3-mini. The key advancement lies in the model’s ability to interact in a more emotionally nuanced and empathetic manner, which elevates the experience for users, especially in collaborative tasks requiring sensitivity and understanding.

Sam Altman on human-like AI

In a post on X (formerly Twitter), OpenAI CEO Sam Altman expressed his excitement about the new model, describing GPT-4.5 as the first AI system that feels like interacting with a human-like entity. "I’ve had several moments where I’ve sat back in my chair and been astonished at getting actually good advice from an AI," Altman noted. He also explained that while the GPT-4.5 AI model isn’t focused on surpassing existing benchmarks, it brings a "different kind of intelligence" that reflects the ongoing technology innovation in AI.

GPT-4.5’s development combines traditional supervised fine-tuning with advanced reinforcement learning from human feedback. These methods, combined with refined Natural Language Processing (NLP) techniques, have allowed OpenAI to create a model that reduces hallucinations while also improving its ability to collaborate effectively in various domains, from creative writing to technical problem-solving.

Availability