OpenAI has introduced GPT-4.5, its latest and most advanced chat model yet. Internally known as Orion, GPT-4.5 is being hailed as the "best model for chat" by OpenAI, offering enhanced pattern recognition, creative insight generation, and emotional intelligence. This model is initially available as a research preview for ChatGPT Pro subscribers and developers worldwide, with plans to expand access to more users in the coming weeks.
Improved Accuracy and Emotional Intelligence: GPT-4.5 reduces AI "hallucinations" (instances of generating false information) to 37.1%, down from 61.8% in GPT-4o. It also demonstrates greater emotional intelligence, making interactions feel more natural and intuitive.
Enhanced User Intent Comprehension: The model is designed to better understand user prompts, providing more accurate and relevant responses. This capability is particularly beneficial for tasks like writing, coding, and problem-solving.
Support for File and Image Uploads: GPT-4.5 allows users to upload files and images, enhancing its utility for projects involving multimedia content. However, it does not yet support advanced voice or video capabilities.
TBC's Insider:
Initial Availability: Currently available to ChatGPT Pro subscribers and developers.
Expansion Plans: Access will be extended to ChatGPT Plus and Team users next week, followed by Enterprise and Edu users the week after.
TBC's Insider:
Despite its impressive capabilities, GPT-4.5 faces challenges due to its massive size and compute-intensive nature. OpenAI CEO Sam Altman described it as a "giant, expensive model," highlighting the need for significant GPU resources. The company plans to add tens of thousands of GPUs to support its rollout.