Premium

Microsoft prepares server capacity as launch of OpenAI’s new GPT models nears: Report

While OpenAI’s GPT-4.5 might be launched early next week, GPT-5 will likely be released in May this year, as per the report.

Microsoft recently laid off people in sales and gaming.

Kalyanaraman, currently CTO, Energy Industry, Asia at Microsoft, will be the first CEO of the ANRF, which, for the time being, is being headed by Secretary of Department of Science and Technology. (File photo)

Microsoft is reportedly making arrangements within its data centres ahead of OpenAI’s highly anticipated launch of the GPT-4.5 and GPT-5 models.

The tech giant is setting up servers to host OpenAI’s upcoming frontier AI models after CEO Sam Altman promised that GPT-4.5 will arrive within a few weeks, according to a report by The Verge.

Since 2019, Microsoft has had arrangements with OpenAI that gave the former the exclusive right to build new computing infrastructure for the latter. However, the agreement between the two tech companies was revised last month to include “changes to the exclusivity on new capacity, moving to a model where Microsoft has a right of first refusal (ROFR).”

While GPT-4.5 can be expected to be launched as early as next week, GPT-5 will likely be released in May this year, as per the report.

GPT-4.5, which has internally been referred to as Orion, is OpenAI’s upgraded model after GPT-4o. It will reportedly be the last non-reasoning AI model in the company’s GPT series.

On the other hand, GPT-5 has been touted to be the most advanced AI model that will integrate other technologies developed by OpenAI. GPT-5 will reportedly be a combination of OpenAI’s o-series and GPT-series models.

Also Read | OpenAI relaxes censorship in AI models, embraces ‘free expression’ in policy shift

In December last year, the company previewed its most advanced reasoning model, o3, and released a mini version for free last month. However, it is unlikely that o3 will be released as a standalone product.

Story continues below this ad

Instead, OpenAI plans to fold all its large language models (LLMs) into a significantly improved model that could be termed as artificial general intelligence or AGI.

“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,” Altman had said in a post on X last week.

While free users of ChatGPT will get access to GPT-5, they will be limited to the “standard intelligence setting, subject to abuse thresholds.”

Additionally, ChatGPT Plus subscribers ($20 per month) will be able to run GPT-5 at a higher level of intelligence, while Pro subscribers ($200 per month) will have access to an even higher level of intelligence. GPT-5 will incorporate “voice, canvas, search, deep research, and more,” Altman said.