OpenAI CEO Sam Altman stated that the corporate was pressured to stagger the rollout of its latest mannequin, GPT-4.5, as a result of OpenAI is “out of GPUs.”
In a post on X, Altman stated that GPT-4.5, which he described as “big” and “costly,” would require “tens of 1000’s” extra GPUs earlier than further ChatGPT customers can acquire entry. GPT-4.5 will come first to subscribers to ChatGPT Professional beginning Thursday, adopted by ChatGPT Plus clients subsequent week.
Maybe partly as a result of its monumental dimension, GPT-4.5 is wildly costly. OpenAI is charging $75 per million tokens (~750,000 phrases) fed into the mannequin and $150 per million tokens generated by the mannequin. That’s 30x the enter price and 15x the output price of OpenAI’s workhorse GPT-4o mannequin.
GPT 4.5 pricing is unhinged. If this doesn’t have monumental fashions scent, I can be dissatisfied pic.twitter.com/1kK5LPN9GH
— Casper Hansen (@casper_hansen_) February 27, 2025
“We’ve been rising loads and are out of GPUs,” Altman wrote. “We’ll add tens of 1000’s of GPUs subsequent week and roll it out to the Plus tier then […] This isn’t how we need to function, however it’s arduous to completely predict progress surges that result in GPU shortages.”
Altman has beforehand stated {that a} lack of computing capability is delaying the corporate’s merchandise. OpenAI hopes to fight this within the coming years by growing its personal AI chips, and by constructing a large community of datacenters.