OpenAI is preparing to launch its next-generation GPT-5 model as early as August, following months of testing and infrastructure adjustments.
The upcoming model was initially expected in late May, with Microsoft engineers reportedly preparing server capacity in advance. However, delays pushed the rollout to early August. OpenAI CEO Sam Altman recently confirmed on X that GPT-5 is coming soon, and even gave a glimpse of its capabilities during a podcast appearance with Theo Von. Altman described an impressive moment where GPT-5 instantly answered a question he couldn’t figure out, calling it a “here it is moment” and admitting he felt “useless” compared to the model’s intelligence.
According to sources familiar with OpenAI’s plans, GPT-5 will include integrated o3-level reasoning capabilities and will unify both the GPT and o-series models into a single, more powerful system. The model will be available in three versions: the full GPT-5, a GPT-5 mini version, and a GPT-5 nano version. While the full and mini models will be accessible through ChatGPT and OpenAI’s API, the nano version will be limited to API access only. Despite its powerful capabilities, Altman has indicated that GPT-5 may not reach its “gold level” performance immediately, with improvements expected over the following months.
Before the release of GPT-5, OpenAI is expected to launch a new open-weight language model by the end of July. This model, described as similar to o3 mini, will include advanced reasoning features and marks the first time OpenAI has released an open-weight model since GPT-2 in 2019. It will be available through platforms like Azure, Hugging Face, and other major cloud providers.
Meanwhile, Microsoft, OpenAI’s primary infrastructure partner, is once again under scrutiny over serious security issues. A major vulnerability in on-premises versions of SharePoint was exploited by Chinese-affiliated hacking groups, compromising more than 50 organizations, including the U.S. nuclear weapons agency. Microsoft issued emergency patches between July 20 and 22 after the flaw was detected on July 18. The SharePoint Online version in Microsoft 365 was not affected. The exploited vulnerability appears to stem from two bugs demonstrated during the Pwn2Own hacking contest earlier this year.
In a separate but related development, a report from ProPublica revealed that Microsoft had been using engineers based in China to help maintain U.S. Department of Defense systems. The practice raised national security concerns, especially after U.S. intelligence agencies warned that China remains one of the most persistent cyber threats. In response, Microsoft has now restricted access to its government cloud systems to U.S.-based engineers only, with a strict “no exceptions” policy enforced starting July 23. The move follows growing political pressure and increased scrutiny over Microsoft’s security posture, especially after a recent scathing review from the U.S. Cyber Safety Review Board.
If this article helped you, please consider supporting our work. Every small contribution keeps Abijita.com independent and running.
