Photo credit: www.phonearena.com

OpenAI Faces GPU Shortage, Delaying GPT-4.5 Release

OpenAI is currently experiencing a significant delay in the rollout of its new GPT-4.5 model due to a shortage of Graphics Processing Units (GPUs). These chips are essential for training complex AI models, handling vast quantities of data, and executing intricate calculations. Unlike Central Processing Units (CPUs), which process data sequentially, GPUs are designed to manage parallel processing, making them more efficient for the matrix operations required in artificial intelligence.

According to a tweet from OpenAI’s CEO, Sam Altman, there is hope as hundreds of thousands of GPUs are expected to be delivered soon. He noted that these resources would quickly be integrated into operations. Presently, the new GPT-4.5 model is accessible only to users subscribing to the Pro version of ChatGPT, priced at $200 per month. However, once the new GPUs are available, users of the ChatGPT Plus subscription, at $20 per month, will also enjoy the advantages of this latest model.

Unlike easily purchased consumer electronics, acquiring GPUs in bulk is not a straightforward process. OpenAI is now contemplating the development of its own chips to mitigate reliance on NVIDIA, whose inventory levels have been unpredictable.

In his commentary on GPT-4.5, Altman clarified that while the model is innovative, it is not designed to outperform existing reasoning models or set new records in benchmarking. He expressed excitement about its unique capabilities, stating, “it’s a different kind of intelligence and there’s a magic to it I haven’t felt before. Really excited for people to try it!”

However, users should prepare for increased costs associated with the new model. Altman highlighted that GPT-4.5 operates at a price of $75 per million input tokens and $150 per million output tokens. In contrast, the previous model, GPT-4o, is significantly more affordable, charging $2.50 per million input tokens and $10 per million output tokens. Tokens represent units of text, encompassing everything from letters and words to punctuation, and Large Language Models (LLMs) process text by breaking it into these tokens.

The surge in demand for NVIDIA’s GPUs is evident, as the company has seen its market value rise dramatically—an impressive 1,748.96% over the past five years—as AI technologies continue to capture significant attention and investment.

Source
www.phonearena.com