OpenAI rolls out latest model GPT-4.5, GPU shortages delay wider rollout
Early tests suggest OpenAI's GPT‑4.5’s responses feel more “natural”, with a stronger knowledge base and improved ability to understand the user’s intent.
ChatGPT-maker on Friday released a research preview of its latest model, GPT-4.5, a general-purpose large language model (LLM).
According to CEO Sam Altman, OpenAI will initially open the newly launched model to software developers and people with ChatGPT Pro subscriptions. In a post on X, he stated that speaking to the GPT-4.5 model "feels like talking to a thoughtful person".
However, Altman said it is an “expensive model”, which led the company to face hardware constraints due to GPU (Graphic Processing Units) shortages.
Early tests suggest GPT‑4.5’s responses feel more “natural”, with a stronger knowledge base and improved ability to understand the user’s intent. It stated that its enhanced “EQ” (Emotional Quotient) makes it particularly useful for tasks like writing, coding, and practical problem-solving.
OpenAI also noted that the model is less prone to hallucinations than its previous models.
In a blog post, the company said, “By scaling unsupervised learning, GPT‑4.5 improves its ability to recognise patterns, draw connections, and generate creative insights without reasoning. GPT‑4.5 is an example of scaling unsupervised learning by scaling up compute and data, along with architecture and optimisation innovations. GPT‑4.5 was trained on Microsoft Azure AI supercomputers. The result is a model that has broader knowledge and a deeper understanding of the world, leading to reduced hallucinations and more reliability across a wide range of topics.”
For now, GPT‑4.5 can be accessed by ChatGPT Pro subscribers through the model picker on the web, mobile, and desktop. The company plans to extend availability to Plus and Team users next week, followed by Enterprise and Edu subscribers in the subsequent week.
Current features such as Search, Canvas, and file uploads remain supported, though multimodal functions like Voice Mode, video, and screen sharing are not yet enabled for GPT‑4.5 in the AI chatbot.
Edited by Suman Singh
