28-2-2025 – OpenAI’s chief executive Sam Altman has revealed that the firm’s latest artificial intelligence model, GPT-4.5, will become available to Plus subscribers next week, describing the technology as possessing a conversational quality akin to “talking to a thoughtful person”.
The anticipated launch comes as the AI developer grapples with significant hardware limitations. “We’re out of GPUs,” Altman acknowledged in his social media announcement, referencing the specialised graphics processing units essential for running sophisticated AI systems.
GPT-4.5 is ready!
good news: it is 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.
bad news: it is a giant, expensive model. we…
— Sam Altman (@sama) February 27, 2025
To address these constraints, OpenAI intends to substantially expand its computing infrastructure, with “tens of thousands of GPUs” being added to support the Plus tier deployment, followed by a more substantial expansion of “hundreds of thousands” in the near future.
Altman expressed frustration at the timing challenges, stating: “This isn’t how we want to operate, but it’s hard to perfectly predict growth surges that lead to GPU shortages.”
Whilst the executive temperedv expectations regarding GPT-4.5’s performance on standardised metrics, suggesting it won’t “crush benchmarks”, he emphasised its distinctive qualities. “There’s a magic to it I haven’t felt before,” Altman remarked, describing the technology as representing “a different kind of intelligence”.
This release aligns with OpenAI’s strategic roadmap previously outlined by Altman, which positions GPT-4.5 (known internally as “Orion”) as the company’s “last non-chain-of-thought model”. Looking ahead, the firm plans to consolidate its various offerings under the GPT-5 umbrella, which will incorporate multiple technologies and feature tiered access across free, Plus and Pro subscription levels.
