Microsoft is ramping up its server capacity in preparation for the upcoming launches of OpenAI’s much-anticipated GPT-4.5 and GPT-5 models, according to a recent report. The Verge reported that the tech giant is actively setting up servers within its data centers to support these next-generation AI models.
Since 2019, Microsoft has maintained an exclusive agreement with OpenAI, providing the computing infrastructure for the AI company’s advancements. This arrangement has been updated, shifting to a “right of first refusal” model for new capacity.
Reportedly, GPT-4.5, internally known as Orion, could be released as early as next week, with GPT-5 expected in May of this year. GPT-4.5 is positioned as the final non-reasoning AI model in the GPT series, while GPT-5 is anticipated to be a more advanced model integrating other OpenAI-developed technologies, including a combination of the o-series and GPT-series models.
OpenAI has been working on improving their models. The company previewed its sophisticated reasoning model, o3, in December, and released a streamlined version last month. OpenAI plans to integrate its large language models (LLMs) to create a unified model that can be described as artificial general intelligence (AGI).
Sam Altman, OpenAI’s CEO, stated on X last week, “We want AI to ‘just work’ for you; we realize how complicated our model and product offerings have become. We hate the model picker as much as you do and want to return to magic unified intelligence.” Free users of ChatGPT will have access to GPT-5, but with a “standard intelligence setting” limited by usage thresholds. ChatGPT Plus subscribers ($20 per month) can operate GPT-5 at a higher intelligence level, while Pro subscribers ($200 per month) will have access to the highest intelligence tier. Altman said that GPT-5 will incorporate “voice, canvas, search, deep research, and more.”
