OpenAI releases GPT 4.5 model – but at a high price amid scepticism about its value

OpenAI co-founder Sam Altman claims GPT 4.5 is the first AI model that can offer ‘good advice’

OpenAI has released its latest AI model, GPT 4.5, a “research preview” that is also the organisation’s most expensive offering due to the limited availability of GPUs.

While not a “frontier model” and more of a step upgrade, it is nevertheless OpenAI’s biggest large language model (LLM) to date and ought to provide more natural conversations and be better at avoiding mistakes.

Indeed, it is, claimed co-founder and CEO Sam Altman “the first model that feels like talking to a thoughtful person to me. I have had several moments where I’ve sat back in my chair and been astonished at getting actually good advice from an AI” in a posting on X, the social network formally known as Twitter.

However, he added, AI friendship comes at a price: “It is a giant, expensive model. We really wanted to launch it to “plus” and “pro” [users] at the same time, but we’ve been growing a lot and are out of GPUs.

“We will add tens of thousands of GPUs next week and roll it out to the plus tier then. (Hundreds of thousands coming soon, and I’m pretty sure y’all will use every one we can rack up.) This isn’t how we want to operate, but it’s hard to perfectly predict growth surges that lead to GPU shortages.”

At the moment, though, it lacks features like a voice mode – shown off recently by Grok – and video support.

AI specialists privileged to have tested previews of GPT 4.5 had mixed early reviews. Ethan Mollick, an associate professor at Wharton, University of Pennsylvania, described it as “odd and interesting”. He claimed that it can “write beautifully” and can be “very creative”, but is also “occasionally oddly lazy on complex projects”.

Filip Piekniewski, a researcher working on computer vision and AI, described it as “underwhelming and expensive”.

Gary Marcus, a scientist involved in two AI companies was also sceptical. “Can I just say that spending $500 billion dollars to serve up models like GPT 4.5 is going to be a colossal flop?” he asked. He added that while the company may have a well-known brand name the company is ultimately in “serious trouble”.

Among other criticisms were its huge expensive – especially GPT 4.5 – has no decisive advantage over rivals despite the level of resources invested, and can offer no killer app, as such. Moreover, a number of top people have left, many to pursue their own rivals, and the company’s burn rate is too high to sustain it to take off, with Microsoft appearing to scale back its commitments, not just to the company, but to its once ambitious approach to AI.